Stimulation pattern control method, control system, electronic device, and medium

Abstract

The application discloses a stimulation mode control method, a control system, an electronic device and a medium. The control method comprises the following steps: constructing a specific data set of a measured physiological signal; judging whether a to-be-measured physiological signal belongs to the specific data set; when the judgment result is "yes", starting stimulation and selecting a matched stimulation mode, so that the stimulation mode is accurate and effective regulation of the stimulation mode is realized, and the accuracy of the stimulation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of time series data processing, and in particular to a stimulation mode control method, a control system, an electronic device and a medium. Background Art

[0002] Neurostimulation technology involves surgically implanting electrodes in specific areas of the brain or spinal cord. This electrical stimulation modulates the activity of relevant neurons, thereby treating neurological disorders. Compared to traditional invasive surgery, neurostimulation offers advantages such as relative safety, reversibility, and post-operative adjustment. It has demonstrated significant therapeutic effects in treating neurological disorders such as epilepsy and Parkinson's disease. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a control method, a control system, an electronic device and a medium for a stimulation mode.

[0004] The technical solution adopted by the present invention to solve its technical problem is:

[0005] In a first aspect, the method for controlling the stimulation mode includes: constructing a specific data set of measured physiological signals; determining whether the physiological signal to be measured belongs to the specific data set; and when the determination result is "yes", starting stimulation and selecting a matching stimulation mode.

[0006] Furthermore, the physiological signal is detected using multiple leads; the specific data set includes: the propagation mode of the measured physiological signal and its occurrence probability; wherein the propagation mode includes: at least one of the number of warning channels and warning propagation timing of the physiological signal.

[0007] Furthermore, the determination of whether the physiological signal to be measured belongs to the specific data set includes: when the number of warning channels of the physiological signal to be measured is greater than or equal to a set threshold of the number of warning channels, the judgment result is "yes"; and / or when the warning propagation timing of the physiological signal to be measured meets the set threshold of the matching warning propagation timing, the judgment result is "yes".

[0008] Furthermore, the priority of determining the number of warning channels is higher than the timing of warning transmission.

[0009] Furthermore, the matching priorities of the stimulation patterns are, in order, the difference in the number of warning channels, the matching similarity of the warning propagation sequence, and the occurrence probability of the propagation pattern.

[0010] Furthermore, the stimulation mode includes stimulation parameters; the stimulation parameters include at least one of the intensity, pulse width, frequency, and charge density of the stimulation waveform.

[0011] In a second aspect, the present invention provides a stimulation mode control system for running the above-mentioned control method, including: a host computer, which is used to construct the specific data set; a slave computer, which is used to store the specific data set and determine whether the physiological signal to be measured belongs to the specific data set, so as to select a matching stimulation mode.

[0012] Furthermore, the lower computer includes: an acquisition module, a matching module, an early warning module, a judgment module and an alarm module; when the judgment result is "no", the lower computer controls the alarm module to issue an alarm; the upper computer includes: a data loading module, a calculation module and a setting module.

[0013] In a third aspect, the present invention provides an electronic device comprising: a processor and a memory, wherein the memory is used to store machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory are communicatively connected, and the above-mentioned control method is executed when the machine-readable instructions are executed by the processor.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, and the computer program executes the above-mentioned control method when executed by a processor.

[0015] The beneficial effect of the present invention is that the control method and control system of the stimulation mode of the present invention can obtain the early warning results of real-time timing data of multiple channels or leads through the early warning algorithm, comprehensively judge whether to implement stimulation based on the early warning results, and select a matching stimulation mode for stimulation, which can improve the accuracy of stimulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be further described below with reference to the accompanying drawings and examples.

[0017] Figure 1 It is a workflow diagram of the control method of the stimulation mode of the present invention.

[0018] Figure 2 is a schematic diagram of the propagation matrix of the present invention.

[0019] Figure 3 It is a schematic diagram of the clustering results of the similarity matrix of the present invention.

[0020] Figure 4 It is a schematic diagram of the first propagation mode of the present invention and its occurrence probability.

[0021] Figure 5 Schematic diagram of the second propagation mode and its occurrence probability of the present invention.

[0022] Figure 6 Schematic diagram of the third propagation mode and its occurrence probability of the present invention.

[0023] Figure 7 It is a structural diagram of the control system of the stimulation mode of the present invention.

[0024] Figure 8 It is a schematic diagram of the processing results of the electromyographic signal by the clustering algorithm of the present invention.

[0025] Figure 9 It is a schematic diagram of the processing results of epilepsy signals by the clustering algorithm of the present invention.

[0026] Figure 10 This is a feature comparison diagram of the electromyographic signal and the epileptic signal of the present invention. DETAILED DESCRIPTION

[0027] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0028] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be internal communication between two components. In addition, the features defined as "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise specified, "plurality" means two or more. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0029] In this case, the physiological signals, such as but not limited to EEG signals, can be matched to different stimulation modes through the control system to adapt to various neurological diseases. The control method of the stimulation mode in this case is now specifically described using EEG signals of epilepsy as an example.

[0030] Epilepsy states can generally be divided into four phases: interictal, preictal, ictal, and postictal. The interictal phase represents the patient's normal EEG signals, the preictal phase represents the period before the onset of the disease, the ictal phase represents the EEG signals during an epileptic seizure, and the postictal phase represents the period after the onset of the disease. Because preictal EEG signals are more active than normal EEG signals, these signals can be used to provide early warning of epileptic seizures.

[0031] like Figure 1As shown, the present invention provides a method for controlling a stimulation mode, which mainly includes the following steps: constructing a specific data set of measured physiological signals; judging whether the physiological signal to be measured belongs to the specific data set; when the judgment result is "yes", starting the stimulation and selecting a matching stimulation mode; when the judgment result is "no", issuing a warning or disabling the stimulation. Of course, the matching stimulation mode can also be selected by manually operating the control system.

[0032] Optionally, the physiological signal is detected using multiple channels; the specific data set includes: the propagation mode of the measured physiological signal and the probability of its occurrence; the propagation mode includes: at least one of the number of warning channels of the physiological signal and the warning propagation timing. The specific operation process of the control method of the stimulation mode is as follows: S1: obtain a number of propagation modes and the probability of occurrence of each propagation mode to form a specific data set. S2: select at least one propagation mode in the specific data set and set a one-to-one corresponding stimulation mode. S3: multiple channels simultaneously output multiple warning results, and judge whether to start stimulation based on the multiple warning results; if the judgment result is "yes", use the stimulation mode to stimulate the target object.

[0033] It should be noted that when monitoring EEG signals, multiple electrodes (i.e., multiple channels or leads) are typically installed on the subject's brain. These electrodes can be placed at different locations on the subject's brain to monitor EEG signals, with one channel corresponding to one electrode. When the subject is in the preictal stage, the EEG signals detected by the multiple channels change in a specific order. This number of channels can be understood as the number of warning channels, and the order in which these channel positions change can be understood as the epilepsy propagation sequence. Together, these two form the epilepsy propagation pattern. For example, there are five channels, designated Channel 1, Channel 2, Channel 3, Channel 4, and Channel 5, each corresponding to a different location in the subject's brain. When the subject is in the preictal stage, the EEG signals detected by the five channels interact with each other. For example, Channel 1 first detects preictal EEG signals. Over time, Channels 2, 3, 4, and 5 subsequently detect preictal EEG signals. In this case, the epilepsy propagation sequence is "Channel 1 → Channel 2 → Channel 3 → Channel 4 → Channel 5." Of course, the epileptic seizure propagation sequence can also be "channel one → channel three → channel two → channel four → channel five" or "channel one → channel three → channel four → channel two → channel five" and so on. In other words, the number of warning channels for epilepsy can reflect the number of channels of EEG signals in the early stage of an epileptic seizure; the warning propagation sequence of epilepsy can reflect the channel order of EEG signal changes in the early stage of an epileptic seizure. The measured EEG data is sent to the host computer, and the number of warning channels (i.e., the number of channels that generate warnings) and their probability of occurrence or the warning propagation sequence (i.e., the propagation sequence that generates warnings) and their probability of occurrence are used to construct a specific data set.

[0034] Each step is described in detail below.

[0035] Step S1: Acquire several propagation modes and the probability of occurrence of each propagation mode to form a specific data set.

[0036] This step can acquire several propagation modes and the probability of each propagation mode occurring. This can be done automatically or manually. Manual acquisition, for example, involves directly storing existing propagation modes in a host computer. Automatic acquisition, for example, involves utilizing historical data from the pre-epilepsy period of the target subject and processing it using a clustering algorithm to obtain several propagation modes and the probability of each propagation mode occurring. The historical data includes EEG data from different locations monitored by multiple channels. The clustering algorithm, for example, can be a propagation matrix similarity clustering algorithm. Processing the historical data using the clustering algorithm specifically includes the following steps.

[0037] S11. Preprocessing: Perform bandpass filtering on the historical data monitored by all channels to filter out the historical data interval in the early stage of epileptic seizure. Then, use a sliding window with a width of t1 to sequentially take out multiple segments of the historical data interval, and perform root mean square (RMS) processing on each segment to make the waveform of the historical data smoother.

[0038] S12, Mutual convolution: Convolution is the result of multiplying two variables within a certain range and then summing them. Multiple channels are arranged in pairs, and a sliding window with a window width of t2 and a step of Δt is used to select data fragments from the data monitored by the two channels in sequence. Assume that both channels can select M data fragments, which are recorded as M a and M b , the data segment M a and M b For example, M a1 With M b1 、M b2 、M b3 ,...M bj Perform convolution operations respectively, M a2 With M b1 、M b2 、M b3 ,...M bj Perform convolution operations respectively, and so on, M ai With M b1 、M b2 、M b3 ,...M bj Convolution operations are performed separately, so that a total of M×M convolution results can be obtained by performing convolution operations between the data of the two channels.

[0039] S13. Link confirmation: If the maximum value among the M×M convolution results is greater than the threshold X, it is considered that there is a link relationship between the two channels; otherwise, it is considered that there is no link relationship between the two channels.

[0040] S14, time difference calculation: Get the multiple groups of channels with link relationship obtained in step S13, and calculate the time difference between the two data segments corresponding to the maximum value of the convolution result by combining the number of sliding windows and the step Δt. For example, the maximum value in the convolution result is the value of M a1 and M b3 Convolution, but the data segment M a1 and M b3 The corresponding moments are different, so it is necessary to calculate the data segment M a1 and M b3 The time difference between.

[0041] S15, propagation matrix: Assume that the number of channels is N, and repeat steps S11 to S14 to obtain T N×N dimensional propagation matrices (such as Figure 2 (as shown in the figure), the values ​​in the propagation matrix can be specific time differences or binarized values. The horizontal axis of the propagation matrix represents the jth channel (1≤i≤N), and the vertical axis represents the ith channel (1≤j≤N). A non-zero value in the propagation matrix indicates a link between the jth and ith channels. A value of "1" in the propagation matrix indicates that the jth channel is propagated before the i-th channel.

[0042] S16. Propagation Matrix Similarity: Each propagation matrix is ​​vectorized to obtain the vector corresponding to the propagation matrix, and then the similarity between each pair of vectors is calculated. The similarity calculation method can be, for example, the Pearson correlation coefficient, Euclidean distance, etc., which can be selected according to the actual situation.

[0043] S17. Similarity matrix: Using the similarities between T propagation matrices as elements, a T×T dimensional similarity matrix can be obtained.

[0044] S18. Similarity matrix clustering: By clustering similarity matrices, we can obtain several stereotyped propagation patterns and the probability of each propagation pattern. For example, Figure 3 As shown in , K1, K2, and K3 represent the frequencies of propagation mode 1, propagation mode 2, and propagation mode 3, respectively. Each propagation mode represents the linkage relationship between the data of N channels. Based on the frequencies of different propagation modes, the probability of each propagation mode is calculated. Figures 4 to 6As shown, the probability of propagation mode 1 occurring is K1 / (K1+K2+K3), the probability of propagation mode 2 occurring is K2 / (K1+K2+K3), and the probability of propagation mode 3 occurring is K3 / (K1+K2+K3). Furthermore, the start time corresponding to each propagation mode is also different. For example, the start time corresponding to propagation mode 1 is 0-12ms, the start time corresponding to propagation mode 2 is 0-70ms, and the start time corresponding to propagation mode 3 is 0-7ms, indicating that the time of occurrence of each propagation mode is also different.

[0045] The propagation modes and the probability of each propagation mode obtained through the above steps are combined into a specific data set and stored in the host computer. It is important to note that the user can select, edit, create, and save the propagation modes, stimulation modes, or stimulation effect parameters stored in the host computer. If the user believes that the existing propagation mode in the host computer does not meet the requirements, they can create or edit a new propagation mode. After creating or editing, the host computer can automatically recalculate the probability of each propagation mode based on the new propagation mode.

[0046] Step S2: selecting at least one propagation mode in a specific data set and setting a corresponding stimulation mode.

[0047] It should be noted that the user selects at least one of the propagation modes displayed on the host computer and sets the corresponding stimulation mode based on the selected propagation mode. For example, the stimulation waveform intensity, pulse width, frequency, charge density and other parameters (i.e., stimulation parameters) can be set. After setting, the user needs to click the "Effect" button to transmit the selected propagation mode and corresponding stimulation mode to the slave computer or the internal device.

[0048] Step S3: Multiple channels simultaneously output multiple warning results. Based on the multiple warning results, a determination is made as to whether stimulation should be initiated. If the determination is "yes," stimulation is performed on the target subject using the stimulation pattern. Specifically, determining whether the measured physiological signal belongs to a specific data set includes: determining "yes" when the number of warning channels for the measured physiological signal is greater than or equal to a set threshold for the number of warning channels; and / or determining "yes" when the warning propagation timing of the measured physiological signal matches a set value for the warning propagation timing.

[0049] Preferably, the judgment priority of the number of warning channels is higher than the warning propagation timing, that is, when the number of warning channels is greater than the set threshold of the number of warning channels, the judgment result must be "yes", and the judgment process at this time no longer considers whether the warning propagation timing matches.

[0050] Preferably, the stimulation pattern matching priority is in the order of difference in the number of warning channels, matching degree of warning propagation sequence, and probability of propagation pattern occurrence. When selecting a matching stimulation pattern, the epilepsy propagation pattern is compared with the corresponding propagation pattern of the stimulation pattern in the specific data set. Priority is given to the stimulation pattern with the smallest difference in the number of warning channels, followed by the highest matching degree of warning propagation sequence, and finally the stimulation pattern with the highest probability of occurrence.

[0051] It should be noted that when multiple channels simultaneously output multiple warning results, an epilepsy warning algorithm can be used to determine whether a warning is activated or not based on the real-time time series data collected by each channel. The output warning result may include whether a warning is activated. If the warning result is activated, the corresponding number of warning channels and the warning transmission time sequence are recorded. There are many epilepsy warning algorithms depending on the actual situation. The following steps of an epilepsy warning algorithm include: T1: Preprocessing the raw EEG data collected in real time from multiple channels; T2: Calculating the signal features of the EEG data; T3: Classifying the EEG data based on the signal features and outputting the warning result.

[0052] It should be noted that the preprocessing in step T1 includes noise reduction, downsampling and multi-window division. Multi-window division can be performed in a manner where the windows partially overlap or do not overlap. The purpose of multi-window division is to select a sequence segment of EEG data each time. The characteristic signal in step T2 is, for example, a zero-crossing coefficient, where the zero-crossing coefficient is a mapping of the zero-crossing rate or a mapping of the number of zero-crossings; the mapping is a mapping function with a positive correlation or a negative correlation; and the mapping function is linear or nonlinear. The zero-crossing coefficient can reflect the frequency of the data value crossing zero in the sequence segment of the EEG signal. According to the zero-crossing coefficient of the EEG data, normal EEG data and EEG data before an epileptic seizure can be distinguished. For example, the zero-crossing calculation formula can be C = 1-sqrt(num{x(1:N-1).*x(2:N)<0} / (N-1)), where N represents the number of points in the sequence segment to be processed, x(1:N-1) represents the array of the first N-1 points in the sequence segment, x(2:N) represents the array of the last N-1 points in the sequence segment, x(1:N-1).*x(2:N) represents the point-to-point multiplication between the two arrays, and num{x(1:N-1).*x(2:N)<0} / (N-1) represents the probability that the result of the point-to-point multiplication is less than 0, i.e., the zero-crossing rate. The zero-crossing rate is then mapped to a zero-crossing coefficient in the range of 0-1 using 1-sqrt(num{x(1:N-1).*x(2:N)<0} / (N-1)). Under this calculation formula, the larger the zero-crossing coefficient, the lower the oscillation activity of the EEG data, and the smaller the zero-crossing coefficient, the higher the oscillation activity of the EEG data. The zero-crossing coefficient of the sequence segment of the EEG data monitored by each channel is calculated, and each zero-crossing coefficient is input into the trained classifier. The classifier can output the classification result of each sequence segment. For example, the classifier outputs "1" to indicate that the EEG signal segment is in the pre-epileptic state, and outputs "0" to indicate that the EEG signal segment is in a normal state. When the sequence segment of the EEG data in a channel has Y consecutive classification results of "1", it is considered that the target object is in the pre-epileptic state and an early warning needs to be activated; otherwise, no early warning is issued. If the judgment result is to activate the early warning, it is necessary to record the time corresponding to the EEG data with the classification result of "1" (i.e., the early warning time) and the channel number (i.e., the early warning channel).

[0053] In the present invention, judging whether to start stimulation according to multiple warning results specifically includes: S31: sorting the warning moments to obtain the actual warning propagation sequence; S32: if the number of warning channels N is SZ is greater than or equal to the first threshold N, then directly start the stimulation; S32: If the number of warning channels N SZ If it is less than the first threshold N, the actual warning propagation timing is matched with the propagation mode. If the matching result is that the actual warning propagation timing is a subset of the propagation mode, the stimulation is started again.

[0054] It should be noted that the recorded multiple warning moments are sorted in ascending or descending order (consistent with the time sequence of the propagation mode) to generate the actual warning propagation time sequence, and the number of warning channels is recorded at the same time. SZ If the number of warning channels N is greater than or equal to the first threshold value N, the stimulation is directly started, and the stimulation mode preset in step S2 is used. SZ If the value is less than the first threshold N, the actual warning timing is matched with the propagation pattern. If the matching result shows that the actual warning propagation timing is a subset of the propagation pattern, stimulation is initiated. The propagation pattern is essentially a time sequence, which is spliced ​​together by the moments of multiple channels. When the actual warning propagation timing is a subset of the propagation pattern, it is considered that stimulation also needs to be initiated, otherwise no stimulation is performed. For example, if the timing sequence of the propagation pattern is 12345, if the actual warning timing is a subset such as 123 or 234, stimulation also needs to be initiated. During stimulation, the stimulation pattern preset in step S2 is adopted. The user can repeat steps S1 to S3 according to the stimulation effect.

[0055] like Figure 7 As shown, the present invention also provides a control system for a stimulation mode, which runs the control method of the above-mentioned stimulation mode. The system includes a host computer 1 and a slave computer 2. The host computer 1 (such as a remote terminal, a CPU, a cloud server) is used to acquire, save and display a specific data set, and set the stimulation mode. The slave computer 2 (such as a computer, a processor) is used to store the specific data set and determine whether the physiological signal to be measured belongs to the specific data set to select a matching stimulation mode. The slave computer includes: an acquisition module 21, a matching module 22, an early warning module 23, a judgment module 24 and an alarm module 25 (such as a sound prompt, disabling the stimulation control circuit, etc.); when the judgment result is "no", the slave computer 2 controls the alarm module 25 to issue an alarm; the host computer 1 includes: a data loading module 11, a calculation module 12 and a setting module 13.

[0056] Preferably, the control system can also be used for implants, and further includes an internal machine 3 installed in the target subject's body for collecting real-time time series data of multiple channels and initiating stimulation. The upper computer 1 is connected to the internal machine 3, and the lower computer 2 is connected to the internal machine 3.

[0057] Specifically, the acquisition module 21 is used to receive the real-time time series data of the target object collected by the internal device 3. The warning module 23 is used to receive the real-time time series data obtained by the acquisition module 21, and process the real-time time series data with the epilepsy warning algorithm to obtain the actual warning propagation time series and the number of warning channels. The judgment module 24 is used to judge the number of warning channels N SZWhether the value of the first threshold N is greater than or equal to the first threshold N, which can be set according to actual conditions. The matching module 22 is configured to match the actual warning propagation sequence with the propagation pattern. If the matching result shows that the actual warning propagation sequence is a subset of the propagation pattern, the matching result is sent to the in-vivo device 3, which then applies stimulation to the target object.

[0058] Specifically, the data loading module 11 is connected to the calculation module 12, and the calculation module 12 is connected to the setting module 13; the setting module 13 is connected to the matching module 32. The data loading module 11 is used to read and display historical data, as well as the number of channels. The calculation module 12 is used to calculate and display a specific data set. The host computer 1 also has a progress bar function, which can be used to check the calculation process. Because the display interface may be stuck when the amount of calculation is relatively large, if the progress bar is still working, it indicates that the entire operation is still proceeding normally and there is no need to restart the operation. The setting module 13 is used to select the propagation mode, and can also edit, create and save the propagation mode. At the same time, the corresponding stimulation mode can be set according to the selected propagation mode, and the selected propagation mode and stimulation mode are transmitted to the internal machine 3. When the internal machine 3 applies stimulation, it stimulates according to the preset stimulation mode.

[0059] Example

[0060] This embodiment can also use the clustering algorithm in the above control method to process the patient's epileptic signal and electromyographic signal separately to obtain Figures 8 to 10 The results can be used to distinguish epileptic signals from electromyographic signals to avoid misjudgment and misstimulation. Figure 8 It is the original waveform of the electromyographic signal and the waveform after RMS processing. Figure 9 The original waveform of the epileptic signal and the waveform after RMS processing. Comparing the two figures, we can find that the propagation of the electromyographic signal has no time sequence, while the propagation of the epileptic signal has time sequence. Figure 10 Epilepsy signals have fewer channels with high link strength, resulting in concentrated link strength and larger time differences. Myoelectric signals, on the other hand, have more channels with high link strength, dispersed link strength, and smaller time differences. This demonstrates that this method can effectively distinguish epilepsy signals from myoelectric signals, effectively improving the accuracy of epilepsy stimulation and reducing misjudgments.

[0061] The present invention also provides an electronic device comprising a processor and a memory, wherein the memory is used to store machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory are communicatively connected, and when the machine-readable instructions are executed by the processor, the steps of the above-mentioned stimulation mode control method are executed.

[0062] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to execute the steps of the control method of the above-mentioned stimulation pattern. In practical applications, the computer-readable medium may be included in the above-mentioned system or may exist separately. The computer-readable storage medium carries one or more programs, and when one or more programs are executed, the described analysis method is implemented. The computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CDROM), an optical storage device, a magnetic storage device, or any suitable combination of the above, but is not used to limit the scope of protection of this application.

[0063] In summary, the control method and control system of the stimulation mode of the present invention, combined with the propagation mode and early warning algorithm of the measured physiological signal, construct a specific data set of the measured physiological signal, and then determine whether the physiological signal to be measured belongs to the specific data set; when the judgment result is "yes", the stimulation is started and the matching stimulation mode is selected, thereby achieving effective regulation of the stimulation mode, avoiding ineffective stimulation and its side effects, and significantly improving the accuracy and safety of the stimulation. In addition, the present invention can also effectively distinguish between epilepsy signals and electromyographic signals, improve the accuracy of epilepsy warning and epilepsy stimulation, reduce misjudgments, and has high application value.

[0064] With the above-described preferred embodiments of the present invention as a guide, and with reference to the above description, relevant personnel are fully capable of making various changes and modifications without departing from the technical spirit of this invention. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A control system for a stimulation mode, characterized in that: include: A host computer, which is used to construct a specific data set; The specific data set includes: propagation patterns of measured physiological signals and their occurrence probabilities; The propagation mode includes at least one of the number of warning channels of the physiological signal and the warning propagation time sequence; The physiological signal is detected using multiple channels; The process of obtaining the specific data set includes: Preprocessing: Band-pass filtering is performed on the historical data monitored by all channels to screen out the historical data interval in the early stage of epileptic seizure. Then, a sliding window with a width of t1 is used to sequentially extract multiple segments of the historical data interval, and root mean square processing is performed on each segment. Mutual convolution: Multiple channels are arranged and combined in pairs, and a sliding window with a window width of t2 and a step of Δt is used to select data fragments from the data monitored by the two channels in sequence, and a pairwise convolution operation is performed between the data fragments of the two channels; Link confirmation: If the maximum value in the convolution result is greater than the threshold X, it is considered that there is a link relationship between the two channels, otherwise it is considered that there is no link relationship between the two channels; Time difference calculation: obtain multiple groups of channels with linked relationships, combine the number of sliding windows and the step Δt, and calculate the time difference between the two data segments corresponding to the maximum value of the convolution result; Propagation matrix: Assume the number of channels is N, and obtain T N×N-dimensional propagation matrices. The horizontal axis of the propagation matrix represents the jth channel, 1≤j≤N, and the vertical axis of the propagation matrix represents the ith channel, 1≤i≤N. If the value in the propagation matrix is ​​non-zero, it indicates that there is a link relationship between the jth channel and the ith channel. If the value in the propagation matrix is ​​"1", it indicates that the propagation order of the jth channel is before the ith channel. Propagation matrix similarity: Each propagation matrix is ​​vectorized to obtain the vector corresponding to the propagation matrix, and then the similarity between the two vectors is calculated; Similarity matrix: Using the similarity between each of T propagation matrices as elements, a T×T dimensional similarity matrix is ​​obtained; Similarity matrix clustering: cluster similarity matrices to obtain several stereotyped propagation patterns and the probability of each propagation pattern; A lower computer, which is used to store the specific data set and determine whether the physiological signal to be measured belongs to the specific data set, and when the determination result is "yes", start stimulation and select a matching stimulation mode; Determining whether the physiological signal to be measured belongs to the specific data set includes: When the number of warning channels of the physiological signal to be measured is greater than or equal to the set threshold value of the number of warning channels, the judgment result is "yes"; and / or, When the warning propagation timing of the physiological signal to be measured matches the set value of the warning propagation timing, the judgment result is "yes".

2. The control system according to claim 1, wherein: The priority of determining the number of warning channels is higher than the timing of warning transmission.

3. The control system according to claim 1, wherein: The matching priorities of the stimulation patterns are, in order, the difference in the number of warning channels, the matching degree of the warning propagation sequence, and the occurrence probability of the propagation pattern.

4. The control system according to claim 1, wherein: The stimulation pattern includes stimulation parameters; The stimulation parameters include at least one of the intensity, pulse width, frequency, and charge density of the stimulation waveform.

5. The control system according to claim 1, wherein: The lower computer includes: an acquisition module, a matching module, an early warning module, a judgment module and an alarm module; When the judgment result is "no", the lower computer controls the warning module to issue a warning; The host computer includes: a data loading module, a calculation module and a setting module.

Images