ECoG electrode performance monitoring and hardware parameter optimization method, device and equipment

Abstract

The invention discloses an ECoG electrode performance monitoring and hardware parameter optimization method, device and equipment, and relates to the field of medical technology assistion.The method comprises the steps that impedance testing and parameter testing are conducted on an ECoG electrode, and signals of all channels of the ECoG electrode are collected to achieve signal analysis; performing performance state evaluation on the target ECoG electrode based on the signal analysis result and the test result; and determining a hardware problem of the target ECoG electrode based on a signal analysis result, and realizing parameter optimization. According to the method, the electrode state can be evaluated, problems can be traced, and theoretical optimal parameter matching of the electrode can be obtained through calculation.

Description

Technical Field

[0001] The present application relates to the field of medical technology assistance, and specifically to an ECoG electrode performance monitoring and hardware parameter optimization method, device and equipment. Background Art

[0002] Currently, ECoG (Electrocorticography) electrodes are designed to be implanted in the skull for a long time, and transmit the signals collected by the electrodes to an external host computer for algorithm decoding before targeted implementation. Currently, hardware testing of ECoG electrodes generally involves electrochemical testing, impedance testing, quality control testing, etc. before electrode implantation. If the current version of the electrode does not meet the requirements and needs to be improved, the corresponding target size parameters need to be experimentally tried based on a large number of electrodes with different specifications and parameters.

[0003] After the ECoG electrode is implanted in the skull, the performance of the electrode is generally judged based on the signal-to-noise ratio, impedance, etc. The judgment method lacks comprehensiveness, and the technology that reversely judges the current performance status of the electrode from the signal decoding effect in combination with actual application scenarios is rare. Strictly speaking, optimizing and improving the hardware based on the problems exposed by the electrode operation status is more of a qualitative improvement. For example, parameters such as contact size, contact spacing, and number of contacts are not directly calculated based on the accumulation of previous experimental data, but require the next large number of experiments to continue. From all the electrodes used in the next experiment, the best performing electrode is selected as the revised electrode. However, the selected best electrode is only the relatively best among the current batch of electrodes that have been manufactured, not the best electrode that exists in theory. Summary of the invention

[0004] The present application provides an ECoG electrode performance monitoring and hardware parameter optimization method, device and equipment, which can evaluate the electrode status and trace problems, and can also calculate the theoretical optimal electrode parameter combination.

[0005] In a first aspect, an embodiment of the present application provides an ECoG electrode performance monitoring and hardware parameter optimization method, the ECoG electrode performance monitoring and hardware parameter optimization method comprising:

[0006] Conduct impedance and parameter tests on ECoG electrodes, and collect signals from each channel of the ECoG electrodes to perform signal analysis;

[0007] Based on the signal analysis results and test results, the performance status of the target ECoG electrode is evaluated;

[0008] Based on the signal analysis results, the hardware problems of the target ECoG electrode are determined and parameter optimization is achieved.

[0009] In combination with the first aspect, in one implementation,

[0010] The impedance test is to perform impedance test on each channel of the ECoG electrode to obtain the impedance test result:

[0011]

[0012] Where R represents the impedance test result of the ECoG electrode, R mn represents the impedance of the contact of the mth row and nth column of the ECoG electrode, and each contact of the ECoG electrode is a channel, and m is the total number of rows of the ECoG electrode contacts, n is the total number of columns of the ECoG electrode contacts, R ij represents the impedance of the contact at the i-th row and j-th column of the ECoG electrode;

[0013] The parameter test is a size parameter test of the ECoG electrode, including contact size C1, contact row spacing C2, contact column spacing C3, and contact distribution area C4. The size parameter test result is C=[C1, C2, C3, C4].

[0014] In combination with the first aspect, in one implementation,

[0015] The signal analysis includes bad track screening, impedance analysis, signal-to-noise ratio analysis, and decoding;

[0016] The bad track screening is used to evaluate whether the channel of the ECoG electrode is abnormal, so as to evaluate the state of the ECoG electrode channel, wherein when any judgment condition is met, it is determined that the current channel is abnormal in the current signal acquisition time period;

[0017] The judgment conditions include that the data collected at each moment are all 0 values, random abnormal value interference at random moments and the number of occurrences is less than a preset number, the data is clearly segmented and there is random abnormal value interference at segmented moments, the data baseline is lower than the set baseline value and the amplitude is less than the preset amplitude, the random moment specific abnormal value interference and the number of occurrences is greater than the set number, the data baseline shows a decreasing trend and the random moment specific abnormal value interference and the number of occurrences is greater than the set threshold;

[0018] The evaluation of the ECoG electrode channel status specifically includes:

[0019] In the total time range of ECoG electrode signal acquisition, the time proportion of each channel being abnormal is counted. Based on the time proportion of the channel being abnormal, the channels with high abnormality in the ECoG electrodes are eliminated. Specifically:

[0020]

[0021] Among them, τ ij represents the time proportion of abnormality in the channel of row i and column j in the ECoG electrode. When τ ijWhen ≥α, the current channel is judged as a channel with a high degree of abnormality. tT represents the total time range length of ECoG electrode signal acquisition, Δt ij represents the total time when the channel in the i-th row and j-th column of the ECoG electrode is abnormal within the total time range of signal acquisition, and α represents the threshold;

[0022] For the channels retained in the ECoG electrodes, create a channel set U1 and mark the time intervals where abnormalities occur in the retained channels: tA ij =[[t t1 , t t2 ],[t t3 , t t4 ]…[t tp , t tp+1 ]],tA ij Indicates the time interval when the channel in the i-th row and j-th column of the ECoG electrode is abnormal, t tp Indicates the start time of the pth abnormality, t tp+1 Indicates the end time when the pth exception occurs;

[0023] Eliminate the impedance of the channel with abnormally high impedance in R and obtain R1.

[0024] In conjunction with the first aspect, in one implementation, impedance analysis specifically includes:

[0025] The dynamic impedance changes of each channel in U1 are stored to obtain the set R U1 ;

[0026] According to the abnormal time interval marking results of each channel in U1, R U1 The corresponding impedance value fragments of each channel in are removed to obtain R U1t ;

[0027] Calculate R U1t The ratio of the impedance variance of each channel to the impedance average value is used to obtain the impedance variation coefficient matrix

[0028] R U1t The impedance values ​​of each channel are fitted by the least square method to obtain the fitting approximate straight line l for each channel. R1t , calculate l R1t The slope obtains the impedance slope matrix k R1t ;

[0029] Calculate the impedance value of each channel in R1 to l R1t The impedance value offset distance matrix d is obtained. R1t .

[0030] In combination with the first aspect, in one implementation,

[0031] For signal-to-noise ratio analysis, it specifically includes:

[0032] The signals of each channel in U1 are clipped by sliding windows, and the signal-to-noise ratio of each window segment is calculated to form a signal-to-noise ratio matrix SNR1. Each channel in SNR1 contains multiple signal-to-noise ratio values.

[0033] Calculate the ratio of the signal-to-noise ratio variance of each channel in SNR1 to the average signal-to-noise ratio to obtain the signal-to-noise ratio variation coefficient matrix

[0034] Calculate the signal-to-noise ratio of each channel in SNR1, and perform least squares fitting to obtain the slope value of the fitting approximate straight line of each channel, and obtain the signal-to-noise ratio slope matrix k SNR1 ;

[0035] For decoding, specifically including:

[0036] Based on the sliding window segments of each channel in U1, the decoding model is trained to obtain the decoding accuracy matrix S of each sliding window of each channel U1 ;

[0037] Calculate the mean decoding accuracy of all window segments of each channel to obtain the decoding accuracy matrix S U1-mean .

[0038] In combination with the first aspect, in one embodiment, the performance status evaluation of the target ECoG electrode based on the signal analysis result and the test result specifically includes:

[0039] R U1t The signals of each channel in the sliding window are clipped to obtain multiple fragments, and the mean of each fragment is calculated to obtain the set R U1t-mean ;

[0040] R U1t-mean 、SNR1、S U1 The values ​​of each channel in R are transposed and stored in columns, and the transposed R U1t-mean 、SNR1、S U1 Perform concat splicing to obtain a fusion matrix, wherein each row in the fusion matrix represents a time segment, and each time segment corresponds to the performance state of an electrode;

[0041] Based on the experimental data, the performance status of ECoG electrodes at different times is collected and made into status labels. U1 ;

[0042] Each row of the fusion matrix is ​​used as training data, and label U1 For labels, a prediction model for electrode performance status is constructed;

[0043] The signal data generated by the target ECoG electrode in the actual application state is input into the electrode performance state prediction model to obtain the performance state evaluation result of the target ECoG electrode at the current moment.

[0044] In combination with the first aspect, in one implementation, determining the hardware problem of the target ECoG electrode based on the signal analysis result specifically includes:

[0045] Construct the channel position matrix U of U1 1xy , U 1xy The x and y in the table represent the rows and columns of each channel in U1 in U0, where U0 is the set of all channels of the ECoG electrodes;

[0046] based on k R1t d R1t , k SNR1 , S U1-mean , U 1xy Construct 7-channel input data;

[0047] Build a convolutional neural network model U1 , the model U1 The input data shape of the first convolutional layer is (m1,n1,7), where m1 represents the total number of rows of U1 and n1 represents the total number of columns of U1;

[0048] Obtain the signal data generated by the actual application state of the ECoG electrode, as well as the corresponding problem type, and train the convolutional neural network model with the problem type as the label U1 ;

[0049] Based on the trained convolutional neural network model U1 Enables prediction of hardware issues for target ECoG electrodes;

[0050] Among them, the types of problems include the falling off of modifying materials, adhesion of electrodes to brain tissue, failure of sealing causing corrosion of electrodes by body fluids, delamination of the insulating layer of the electrode itself, and poor contact of electrode connections.

[0051] In conjunction with the first aspect, in one implementation, the implementing parameter optimization specifically includes:

[0052] Obtain ECoG electrodes of various specifications and perform signal analysis. For each specification of ECoG electrode, obtain C and S U1-mean , U 1xy ;

[0053] The regression model constructed by training the size parameters of each specification of ECoG electrode as labels;

[0054] S U1-mean and U 1xy Multiply the corresponding positions to get the matrix SU U1 , SU U1 Reshape it into one dimension and concatenate it with the size parameters of the current ECoG electrode. The result is input into the regression model to obtain the parameter recommendation model. C ;

[0055] Get the target ECoG electrode and perform signal analysis to get the SU corresponding to the target ECoG electrode U1 , SU U1 Input to model C , and obtain the optimized size parameters of the target ECoG electrode.

[0056] In a second aspect, an embodiment of the present application provides an ECoG electrode performance monitoring and hardware parameter optimization device, the ECoG electrode performance monitoring and hardware parameter optimization device comprising:

[0057] A signal analysis module, which is used to perform impedance testing and parameter testing on the ECoG electrodes, and collect signals from each channel of the ECoG electrodes to perform signal analysis;

[0058] A status evaluation module, which is used to evaluate the performance status of the target ECoG electrode based on the signal analysis results and the test results;

[0059] The hardware problem analysis and parameter optimization module is used to determine the hardware problem of the target ECoG electrode based on the signal analysis results and to achieve parameter optimization.

[0060] In a third aspect, an embodiment of the present application provides an ECoG electrode performance monitoring and hardware parameter optimization device, wherein the ECoG electrode performance monitoring and hardware parameter optimization device comprises a processor, a memory, and an ECoG electrode performance monitoring and hardware parameter optimization program stored in the memory and executable by the processor, wherein when the ECoG electrode performance monitoring and hardware parameter optimization program is executed by the processor, the steps of the above-mentioned ECoG electrode performance monitoring and hardware parameter optimization method are implemented.

[0061] The beneficial effects brought by the technical solution provided by the embodiments of the present application include:

[0062] Based on artificial intelligence technology, the performance of ECoG electrodes that are in actual application for a long time can be monitored in real time, and the electrode status can be evaluated and problems can be traced according to signal quality, impedance, and decoding accuracy, as well as parameter optimization. Specifically, based on the characteristic matrix corresponding to impedance, signal-to-noise ratio, and channel decoding accuracy, the current performance status of the ECoG electrode can be predicted in real time, and the problems of the ECoG electrode in the application state can be located according to the impedance changes, the dynamic changes of the signal-to-noise ratio, and the decoding accuracy. Based on the decoding accuracy and the parameter configuration of the previous generation of electrodes, the next-generation iterative optimization parameters of the electrodes can be predicted with the help of artificial intelligence technology to determine the theoretically optimal electrode parameter combination. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A flowchart of the ECoG electrode performance monitoring and hardware parameter optimization method of this application;

[0064] Figure 2 This is a schematic diagram of the functional modules of the ECoG electrode performance monitoring and hardware parameter optimization device of this application;

[0065] Figure 3 Schematic diagram of the hardware structure of the ECoG electrode performance monitoring and hardware parameter optimization equipment for this application. DETAILED DESCRIPTION

[0066] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0067] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0068] On the first aspect, the embodiments of the present application provide an ECoG electrode performance monitoring and hardware parameter optimization method. Based on artificial intelligence technology, the performance of ECoG electrodes that are in actual application for a long time can be monitored in real time, and the electrode status can be evaluated and problems can be traced based on signal quality, impedance, and decoding accuracy. The theoretically optimal parameter combination can also be calculated through artificial intelligence technology based on the electrode decoding effect.

[0069] In one embodiment, referring to Figure 1 , Figure 1 This is a flow chart of the ECoG electrode performance monitoring and hardware parameter optimization method of this application. Figure 1As shown, the ECoG electrode performance monitoring and hardware parameter optimization method includes:

[0070] S1: Conduct impedance test and parameter test on ECoG electrodes, and collect signals from each channel of ECoG electrodes to realize signal analysis;

[0071] S2: Based on the signal analysis results and test results, the performance status of the target ECoG electrode is evaluated;

[0072] S3: Determine the hardware problem of the target ECoG electrode based on the signal analysis results and implement parameter optimization.

[0073] It should be noted that the above step S1 is to perform signal analysis on a single ECoG electrode, and steps S2 and S3 are based on the signal analysis results and test results of multiple ECoG electrodes, so as to achieve performance status evaluation of the target ECoG electrode (i.e., the ECoG electrode to be subjected to performance status evaluation, hardware problem prediction, and parameter optimization), as well as hardware problem prediction and parameter optimization.

[0074] Furthermore, in one embodiment, the impedance test is to perform an impedance test on each channel of the ECoG electrode to obtain an impedance test result:

[0075]

[0076] Where R represents the impedance test result of the ECoG electrode, R mn represents the impedance of the contact of the mth row and nth column of the ECoG electrode, and each contact of the ECoG electrode is a channel, and m is the total number of rows of the ECoG electrode contacts, n is the total number of columns of the ECoG electrode contacts, R ij represents the impedance of the contact at the i-th row and j-th column of the ECoG electrode;

[0077] The parameter test is a size parameter test of the ECoG electrode, including contact size C1, contact row spacing C2, contact column spacing C3, and contact distribution area C4. The size parameter test result is C=[C1, C2, C3, C4]. The size parameter test result obtained by the test is stored.

[0078] It should be noted that the impedance test and parameter test are performed on the ECoG electrodes before leaving the factory.

[0079] In this application, when performing signal analysis on each channel signal of the ECoG electrode, it mainly includes bad track screening, impedance analysis, signal-to-noise ratio analysis, and decoding.

[0080] Since the ECoG electrode may show obvious abnormal signals in a certain period of time during signal acquisition, but return to normal in the next period of time, the subsequent analysis of this application is based on the normality of the channel, so it is necessary to evaluate whether each channel is abnormal. This application analyzes and evaluates whether each channel is damaged from different angles based on the signals collected by each channel.

[0081] Specifically, the bad track screening is used to evaluate whether the channel of the ECoG electrode is abnormal, so as to evaluate the state of the ECoG electrode channel, wherein when any determination condition is met, it is determined that the current channel is abnormal in the current signal acquisition time period.

[0082] The judgment conditions include that the data collected at each moment are all 0 values, random outlier interference at random moments and the number of occurrences is less than the preset number (i.e., a small amount of random outlier interference at random moments), the data are obviously segmented and there is random outlier interference at the segmented moments (i.e., the data are obviously segmented and there is random outlier interference at the segmented moments), the data baseline is lower than the set baseline value and the amplitude is less than the preset amplitude (i.e., the data baseline is low and the amplitude is small), there is specific outlier interference at random moments and the number of occurrences is greater than the set number (i.e., multiple random outlier interferences), and the data baseline shows a downward trend while there is specific outlier interference at random moments and the number of occurrences is greater than the set threshold (i.e., baseline reduction and multiple random outlier interferences).

[0083] Among them, the evaluation of the ECoG electrode channel status includes:

[0084] S101: within the total time range of ECoG electrode signal acquisition, count the time proportion of each channel being abnormal, and based on the time proportion of the channel being abnormal, remove the channels with high abnormality in the ECoG electrodes, specifically:

[0085]

[0086] Among them, τ ij represents the time proportion of abnormality in the channel of row i and column j in the ECoG electrode. When τ ij When ≥α, the current channel is judged as a channel with a high degree of abnormality. tT represents the total time range length of ECoG electrode signal acquisition, Δt ij represents the total time when the channel in the i-th row and j-th column of the ECoG electrode is abnormal within the total time range of signal acquisition, α represents a threshold value, and α can be given according to experiments or experience, such as α is 0.1, and is not specifically limited in this application;

[0087] S102: For the channels retained in the ECoG electrodes, create a channel set U1 and mark the time intervals where abnormalities occur in the retained channels: tA ij =[[t t1 , tt2 ],[t t3 , t t4 ]…[t tp , t tp+1 ]]tA ij Indicates the time interval when the channel in the i-th row and j-th column of the ECoG electrode is abnormal, t tp Indicates the start time of the pth abnormality, t tp+1 Indicates the end time when the pth exception occurs;

[0088] S103: Eliminate the impedance of the channel with a high degree of abnormality in R to obtain R1.

[0089] In this application, impedance analysis specifically includes:

[0090] S111: Store the dynamic change of impedance of each channel in U1 to obtain a set R U1 , set R U1 Each subset in represents the impedance change of a channel in U1 over the test time;

[0091] Specifically, the dynamic change of impedance of each channel in U1 is stored, for example:

[0092]

[0093] Among them, R mntT represents the impedance of the channel in the mth row and nth column at time tT. “ / ” represents the eliminated channel. The number of channels without “ / ” is equal to the number of channels in U1, that is, R U1 .

[0094] S112: According to the result of marking the abnormal time interval of each channel in U1, R U1 The corresponding impedance value fragments of each channel in are removed to obtain R U1t ;

[0095] S113: Calculate R U1t The ratio of the impedance variance of each channel to the impedance average value is used to obtain the impedance variation coefficient matrix

[0096] S114: R U1t The impedance values ​​of each channel are fitted by the least square method to obtain the fitting approximate straight line l for each channel. R1t , calculate l R1t The slope obtains the impedance slope matrix k R1t ;

[0097] S115: Calculate the impedance value of each channel in R1 to l R1t The impedance value offset distance matrix d is obtained. R1t.

[0098] In this application, the signal-to-noise ratio analysis specifically includes:

[0099] S121: Perform sliding window clipping on the signals of each channel in U1, and calculate the signal-to-noise ratio of each window segment to form a signal-to-noise ratio matrix SNR1, where each channel in SNR1 contains multiple signal-to-noise ratio values;

[0100] S122: Calculate the ratio of the signal-to-noise ratio variance of each channel in SNR1 to the average signal-to-noise ratio to obtain the signal-to-noise ratio variation coefficient matrix

[0101] S123: Calculate the signal-to-noise ratio of each channel in SNR1, and perform least squares fitting to obtain the slope value of the fitting approximate straight line of each channel, and obtain the signal-to-noise ratio slope matrix k SNR1 .

[0102] In this application, decoding specifically includes:

[0103] S131: Train the decoding model based on the sliding window segments of each channel in U1 to obtain the decoding accuracy matrix S of each sliding window of each channel U1 ;

[0104] S132: Calculate the mean of the decoding accuracy of all window segments of each channel to obtain the decoding accuracy matrix S U1-mean .

[0105] Furthermore, in one embodiment, based on the signal analysis results and the test results, the performance status of the target ECoG electrode is evaluated, specifically including:

[0106] S201: R U1t The signals of each channel in the sliding window are clipped to obtain multiple fragments, and the mean of each fragment is calculated to obtain the set R U1t-mean ; The step size of the sliding window at this step is the same as that in step S121;

[0107] S202: R U1t-mean 、SNR1、S U1 The values ​​of each channel in R are transposed and stored in columns, and the transposed R U1t-mean 、SNR1、S U1 Perform concat splicing to obtain a fusion matrix, wherein each row in the fusion matrix represents a time segment, and each time segment corresponds to the performance state of an electrode;

[0108] concat, that is, the concat function, is used to concatenate multiple strings into a new string;

[0109] S203: Based on the experiment, the performance status of the ECoG electrodes at different times is collected and made into status labels U1 Specifically, a large number of experiments can be conducted to collect the performance status of ECoG electrodes at different times, thereby producing a status label. U1 ;

[0110] S204: Use each row of the fusion matrix as training data, and label U1 For labels, a prediction model for electrode performance status is constructed;

[0111] S205: Input the signal data generated by the target ECoG electrode in the actual application state into the electrode performance state prediction model to obtain the performance state evaluation result of the target ECoG electrode at the current moment. If the electrode performance state prediction model predicts that the target ECoG electrode is in a poor state, the use of the target ECoG electrode needs to be stopped.

[0112] Based on the signal quality, signal decoding accuracy, and impedance changes collected by the ECoG electrode, this application can reversely calculate possible problems with the target ECoG electrode itself, such as falling off of contact modification material, adhesion of clicks to brain tissue, sealing failure leading to corrosion of the electrode by body fluids, delamination of the insulation layer of the electrode itself, and poor contact of the electrode connection.

[0113] Furthermore, in one embodiment, determining the hardware problem of the target ECoG electrode based on the signal analysis result specifically includes:

[0114] S301: Construct the channel position matrix U of U1 1xy , U 1xy The x and y in the table represent the rows and columns of each channel in U1 in U0, where U0 is the set of all channels of the ECoG electrodes;

[0115] S302: Based on k R1t d R1t , k SNR1 , S U1-mean , U 1xy Construct 7-channel input data;

[0116] S303: Build a convolutional neural network model U1 , the model U1 The input data shape of the first convolutional layer is (m1,n1,7), where m1 represents the total number of rows of U1, n1 represents the total number of columns of U1; shape, i.e., shape function, is used to obtain the dimension information of the array;

[0117] The algorithm model constructed can be a traditional machine learning model or a deep learning model. This application takes a convolutional neural network model as an example for explanation;

[0118] S304: Obtain the signal data generated by the actual application state of the ECoG electrode and the corresponding problem type, and train the convolutional neural network model model with the problem type as the label U1 ;

[0119] Specifically, we obtain the output signals of a large number of ECoG electrodes in actual application and the corresponding problem types, so as to train the convolutional neural network model with the problem types as labels. U1 ;

[0120] S305: Based on the trained convolutional neural network model U1 Predict hardware problems of the target ECoG electrode; the problem types include falling off of modification materials, adhesion of electrodes to brain tissue, corrosion of electrodes by body fluids due to seal failure, delamination of the insulating layer of the electrode itself, and poor contact of electrode connections.

[0121] Furthermore, in one embodiment, parameter optimization is implemented, specifically including:

[0122] S311: Obtain ECoG electrodes of various specifications and perform signal analysis. For each specification of ECoG electrode, obtain C and S U1-mean , U 1xy ;

[0123] Specifically, ECoG electrodes of various specifications are obtained for testing, and each ECoG electrode of various specifications has the C and S U1-mean , U 1xy ;

[0124] S312: A regression model constructed by training with the size parameters of each specification of ECoG electrode as labels;

[0125] Specifically, the regression model is trained with the optimized and upgraded size parameters of each specification of ECoG electrode as labels. In practical applications, the regression model can be MultiOutputRegressor, DecisionTreeRegressor, RandomForestRegressor, or RNN network model in the machine learning algorithm.

[0126] S313: S U1-mean and U 1xy Multiply the corresponding positions to get the matrix SU U1 , SU U1Reshape it into one dimension and concatenate it with the size parameters of the current ECoG electrode. The result is input into the regression model to obtain the parameter recommendation model. C ;

[0127] reshape, that is, the reshape function, is used to transform the specified matrix into a matrix of a specific dimension.

[0128] S314: Obtain the target ECoG electrode and perform signal analysis to obtain the SU corresponding to the target ECoG electrode U1 , SU U1 Input to model C , and obtain the optimized size parameters of the target ECoG electrode.

[0129] Specifically, for a target ECoG electrode in actual application state, the corresponding SU is obtained according to the output signal of the electrode. U1 , SU U1 Input to model C , you can get the parameter recommendation model model C The next generation optimized size parameters are recommended.

[0130] The ECoG electrode performance monitoring and hardware parameter optimization method of the embodiment of the present application is based on artificial intelligence technology, and can monitor the performance of ECoG electrodes that are in actual application for a long time in real time, and evaluate the electrode status and trace problems according to signal quality, impedance, and decoding accuracy, as well as optimize parameters. Specifically, based on the characteristic matrix corresponding to impedance, signal-to-noise ratio, and channel decoding accuracy, the current performance status of the ECoG electrode can be predicted in real time, and the problems of the ECoG electrode in the application state can be located according to the impedance changes, the dynamic changes of the signal-to-noise ratio, and the decoding accuracy. Based on the decoding accuracy and the parameter configuration of the previous generation of electrodes, the next generation of iterative optimization parameters of the electrodes can be predicted with the help of artificial intelligence technology, and the theoretically optimal electrode parameter combination can be determined.

[0131] In a second aspect, an embodiment of the present application also provides an ECoG electrode performance monitoring and hardware parameter optimization device.

[0132] In one embodiment, referring to Figure 2 , Figure 2 This is a schematic diagram of the functional modules of the ECoG electrode performance monitoring and hardware parameter optimization device of this application. Figure 2 As shown, the ECoG electrode performance monitoring and hardware parameter optimization device includes: a signal analysis module, a state evaluation module, a hardware problem analysis and parameter optimization module.

[0133] The signal analysis module is used to perform impedance testing and parameter testing on the ECoG electrodes, and to collect signals from each channel of the ECoG electrodes to achieve signal analysis; the status evaluation module is used to evaluate the performance status of the target ECoG electrodes based on the signal analysis results and test results; the hardware problem analysis and parameter optimization module is used to determine the hardware problems of the target ECoG electrodes based on the signal analysis results, and to achieve parameter optimization.

[0134] In a third aspect, an embodiment of the present application provides an ECoG electrode performance monitoring and hardware parameter optimization device, which may be a personal computer (PC), a laptop computer, a server, or other device with data processing capabilities.

[0135] Reference Figure 3 , Figure 3 Schematic diagram of the hardware structure of the ECoG electrode performance monitoring and hardware parameter optimization device involved in the embodiment of the present application. In the embodiment of the present application, the ECoG electrode performance monitoring and hardware parameter optimization device may include a processor, a memory, a communication interface and a communication bus.

[0136] The communication bus may be of any type and is used to interconnect the processor, the memory, and the communication interface.

[0137] The communication interface includes input / output (I / O) interface, physical interface and logical interface, etc., which are used to realize the interconnection of devices inside the ECoG electrode performance monitoring and hardware parameter optimization device, and the interface used to realize the interconnection of the ECoG electrode performance monitoring and hardware parameter optimization device with other devices (such as other computing devices or user devices). The physical interface can be an Ethernet interface, a fiber optic interface, an ATM interface, etc.; the user device can be a display, a keyboard, etc.

[0138] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0139] The processor may be a general-purpose processor, which may call the ECoG electrode performance monitoring and hardware parameter optimization program stored in the memory, and execute the ECoG electrode performance monitoring and hardware parameter optimization method provided in the embodiment of the present application. For example, the general-purpose processor may be a central processing unit (CPU). The method executed when the ECoG electrode performance monitoring and hardware parameter optimization program is called may refer to the various embodiments of the ECoG electrode performance monitoring and hardware parameter optimization method of the present application, which will not be repeated here.

[0140] Those skilled in the art will understand that Figure 3 The hardware structure shown in the figure does not constitute a limitation on the present application, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0141] The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices. The terms "first", "second" and "third" are used to distinguish different objects, etc., and do not represent a sequence, nor do they limit "first", "second" and "third" to different types.

[0142] In the description of the embodiments of the present application, "exemplary", "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary", "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "for example" or "for example" is intended to present related concepts in a specific way.

[0143] In the description of the embodiments of the present application, unless otherwise specified, “ / ” means or, for example, A / B can mean A or B; the “and / or” in the text is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, “multiple” refers to two or more than two.

[0144] In some processes described in the embodiments of the present application, multiple operations or steps that appear in a specific order are included, but it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of the present application or in parallel, and the sequence number of the operation is only used to distinguish the different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0145] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD) as described above, and includes a number of instructions for a terminal device to execute the methods described in each embodiment of the present application.

[0146] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for ECoG electrode performance monitoring and hardware parameter optimization, characterized in that: The ECoG electrode performance monitoring and hardware parameter optimization method includes: Conduct impedance and parameter tests on ECoG electrodes, and collect signals from each channel of the ECoG electrodes to perform signal analysis; Based on the signal analysis results and test results, the performance status of the target ECoG electrode is evaluated; Based on the signal analysis results, the hardware problems of the target ECoG electrode are determined and parameter optimization is achieved.

2. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 1, characterized in that: The impedance test is to perform impedance test on each channel of the ECoG electrode to obtain the impedance test result: Where R represents the impedance test result of the ECoG electrode, R mn represents the impedance of the contact of the mth row and nth column of the ECoG electrode, and each contact of the ECoG electrode is a channel, and m is the total number of rows of the ECoG electrode contacts, n is the total number of columns of the ECoG electrode contacts, R ij represents the impedance of the contact at the i-th row and j-th column of the ECoG electrode; The parameter test is a size parameter test of the ECoG electrode, including contact size C1, contact row spacing C2, contact column spacing C3, and contact distribution area C4. The size parameter test result is C=[C1, C2, C3, C4].

3. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 2, characterized in that: The signal analysis includes bad track screening, impedance analysis, signal-to-noise ratio analysis, and decoding; The bad track screening is used to evaluate whether the channel of the ECoG electrode is abnormal, so as to evaluate the state of the ECoG electrode channel, wherein when any judgment condition is met, it is determined that the current channel is abnormal in the current signal acquisition time period; The judgment conditions include that the data collected at each moment are all 0 values, random abnormal value interference at random moments and the number of occurrences is less than a preset number, the data is clearly segmented and there is random abnormal value interference at segmented moments, the data baseline is lower than the set baseline value and the amplitude is less than the preset amplitude, the random moment specific abnormal value interference and the number of occurrences is greater than the set number, the data baseline shows a decreasing trend and the random moment specific abnormal value interference and the number of occurrences is greater than the set threshold; The evaluation of the ECoG electrode channel status specifically includes: In the total time range of ECoG electrode signal acquisition, the time proportion of each channel being abnormal is counted. Based on the time proportion of the channel being abnormal, the channels with high abnormality in the ECoG electrodes are eliminated. Specifically: Among them, τ ij represents the time proportion of abnormality in the channel of row i and column j in the ECoG electrode. When τ ij When ≥α, the current channel is judged as a channel with a high degree of abnormality. tT represents the total time range length of ECoG electrode signal acquisition, Δt ij represents the total time when the channel in the i-th row and j-th column of the ECoG electrode is abnormal within the total time range of signal acquisition, and α represents the threshold; For the channels retained in the ECoG electrodes, create a channel set U1 and mark the time intervals where abnormalities occur in the retained channels: tA ij =[[t t1 , t t2 ],[t t3 , t t4 ]…[t tp , t tp+1 ]], tA ij Indicates the time interval when the channel in the i-th row and j-th column of the ECoG electrode is abnormal, t tp Indicates the start time of the pth abnormality, t tp+1 Indicates the end time when the pth exception occurs; Eliminate the impedance of the channel with abnormally high impedance in R and obtain R1.

4. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 3, characterized in that: For impedance analysis, specifically: The dynamic impedance changes of each channel in U1 are stored to obtain the set R U1 ; According to the abnormal time interval marking results of each channel in U1, R U1 The corresponding impedance value fragments of each channel in are removed to obtain R U1t ; Calculate R U1t The ratio of the impedance variance of each channel to the impedance average value is used to obtain the impedance variation coefficient matrix R U1t The impedance values ​​of each channel are fitted by the least square method to obtain the fitting approximate straight line l for each channel. R1t , calculate l R1t The slope obtains the impedance slope matrix k R1t ; Calculate the impedance value of each channel in R1 to l R1t The impedance value offset distance matrix d is obtained. R1t .

5. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 4, characterized in that: For signal-to-noise ratio analysis, it specifically includes: The signals of each channel in U1 are clipped by sliding windows, and the signal-to-noise ratio of each window segment is calculated to form a signal-to-noise ratio matrix SNR1. Each channel in SNR1 contains multiple signal-to-noise ratio values. Calculate the ratio of the signal-to-noise ratio variance of each channel in SNR1 to the average signal-to-noise ratio to obtain the signal-to-noise ratio variation coefficient matrix Calculate the signal-to-noise ratio of each channel in SNR1, and perform least squares fitting to obtain the slope value of the fitting approximate straight line of each channel, and obtain the signal-to-noise ratio slope matrix k SNR1 ; For decoding, specifically including: Based on the sliding window segments of each channel in U1, the decoding model is trained to obtain the decoding accuracy matrix S of each sliding window of each channel U1 ; Calculate the mean decoding accuracy of all window segments of each channel to obtain the decoding accuracy matrix S U1-mean .

6. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 5, characterized in that: The performance status evaluation of the target ECoG electrode based on the signal analysis results and the test results specifically includes: R U1t The signals of each channel in the sliding window are clipped to obtain multiple fragments, and the mean of each fragment is calculated to obtain the set R U1t-mean ; R U1t-mean 、SNR1、S U1 The values ​​of each channel in R are transposed and stored in columns, and the transposed R U1t-mean 、SNR1、S U1 Perform concat splicing to obtain a fusion matrix, wherein each row in the fusion matrix represents a time segment, and each time segment corresponds to the performance state of an electrode; Based on the experimental data, the performance status of ECoG electrodes at different times is collected and made into status labels. U1 ; Each row of the fusion matrix is ​​used as training data, and label U1 For labels, a prediction model for electrode performance status is constructed; The signal data generated by the target ECoG electrode in the actual application state is input into the electrode performance state prediction model to obtain the performance state evaluation result of the target ECoG electrode at the current moment.

7. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 5, characterized in that: The hardware problem of the target ECoG electrode is determined based on the signal analysis result, specifically including: Construct the channel position matrix U of U1 1xy , U 1xy The x and y in the table represent the rows and columns of each channel in U1 in U0, where U0 is the set of all channels of the ECoG electrodes; based on k R1t d R1t , k SNR1 , S U1-mean , U 1xy Construct 7-channel input data; Build a convolutional neural network model U1 , the model U1 The input data shape of the first convolutional layer is (m1,n1,7), where m1 represents the total number of rows of U1 and n1 represents the total number of columns of U1; Obtain the signal data generated by the actual application state of the ECoG electrode, as well as the corresponding problem type, and train the convolutional neural network model with the problem type as the label U1 ; Based on the trained convolutional neural network model U1 Enables prediction of hardware issues for target ECoG electrodes; Among them, the types of problems include the falling off of modifying materials, adhesion of electrodes to brain tissue, failure of sealing causing corrosion of electrodes by body fluids, delamination of the insulating layer of the electrode itself, and poor contact of electrode connections.

8. The method for ECoG electrode performance monitoring and hardware parameter optimization according to claim 7, characterized in that: The implementation of parameter optimization specifically includes: Obtain ECoG electrodes of various specifications and perform signal analysis. For each specification of ECoG electrode, obtain C and S U1-mean , U 1xy ; The regression model constructed by training the size parameters of each specification of ECoG electrode as labels; S U1-mean and U 1xy Multiply the corresponding positions to get the matrix SU U1 , SU U1 Reshape it into one dimension and concatenate it with the size parameters of the current ECoG electrode. The result is input into the regression model to obtain the parameter recommendation model. C ; Get the target ECoG electrode and perform signal analysis to get the SU corresponding to the target ECoG electrode U1 , SU U1 Input to model C , and obtain the optimized size parameters of the target ECoG electrode.

9. An ECoG electrode performance monitoring and hardware parameter optimization device, characterized in that: The ECoG electrode performance monitoring and hardware parameter optimization device comprises: A signal analysis module, which is used to perform impedance testing and parameter testing on the ECoG electrodes, and collect signals from each channel of the ECoG electrodes to perform signal analysis; A status evaluation module, which is used to evaluate the performance status of the target ECoG electrode based on the signal analysis results and the test results; The hardware problem analysis and parameter optimization module is used to determine the hardware problem of the target ECoG electrode based on the signal analysis results and to achieve parameter optimization.

10. An ECoG electrode performance monitoring and hardware parameter optimization device, characterized in that: The ECoG electrode performance monitoring and hardware parameter optimization device includes a processor, a memory, and an ECoG electrode performance monitoring and hardware parameter optimization program stored in the memory and executable by the processor, wherein when the ECoG electrode performance monitoring and hardware parameter optimization program is executed by the processor, the steps of the ECoG electrode performance monitoring and hardware parameter optimization method according to any one of claims 1 to 8 are implemented.

Images