Electrical stimulation systems, neuromodulation equipment, epilepsy treatment equipment

Abstract

The present invention belongs to the field of electrophysiological signal detection technology, and specifically relates to an electrical stimulation system, a neural regulation device, and an epilepsy treatment device. The electrical stimulation system includes: a classifier; the input end of the classifier is the input feature of the signal after electrical stimulation, and the output end is the classification result of the effectiveness of electrical stimulation; wherein the input feature includes the proportion of stimulation channels with energy changes before and after electrical stimulation and\or the situation where the stimulation channels cover strongly connected nodes. By associating and binding the stimulation mode with the stimulation effect based on the data features of the EEG signal, the effectiveness of the stimulation mode can be effectively predicted in vitro or before the device is implanted, and the stimulation mode can be adjusted accordingly.

Description

Technical Field

[0001] The present invention belongs to the technical field of physiological electrical signal detection, and specifically relates to an electrical stimulation system, a nerve regulation device, and an epilepsy treatment device. Background Art

[0002] Neuromodulation (such as RNS / DBS) is an important treatment for intractable epilepsy, enabling real-time detection of EEG signals or application of electrical stimulation to the brain according to a fixed pattern. In particular, the process of treating epilepsy with implantable devices involves the following steps: First, precise localization of the epileptogenic zone (SOZ). Some patients undergo stereoelectroencephalography (sEEG) and use electrical cortical stimulation (ECS) to locate the functional area and SOZ. Second, a pre-set stimulation pattern is established. The SOZ is stimulated using either cortical stimulation or deep neural stimulation to determine the stimulation parameters and channels. Third, closed-loop detection and stimulation is implemented using the pre-set stimulation pattern after implantation. In related technologies, neuromodulation therapy devices typically use seizure data to determine the effectiveness of stimulation and adjust the stimulation pattern accordingly. Due to individual variability, seizure frequency can vary across individuals even with the same stimulation pattern. Therefore, establishing a stable correlation between the stimulation pattern and the stimulation effect is difficult. Consequently, stimulation patterns based on historical data are not suitable for online stimulation. Adjusting the stimulation pattern online requires a significant amount of online training. Summary of the Invention

[0003] The present invention provides an electrical stimulation system, a neural regulation device, and an epilepsy treatment device to obtain a stable correlation between the stimulation pattern and the electrical stimulation effect, thereby reducing the training time for online adjustment of the stimulation pattern.

[0004] In order to solve the above technical problems, the present invention provides an electrical stimulation system, including: a classifier; the input end of the classifier is the input feature of the signal after electrical stimulation, and the output end is the classification result of the effectiveness of electrical stimulation; wherein the input feature includes the proportion of stimulation channels with energy changes before and after electrical stimulation and\or the situation where the stimulation channels cover strongly connected nodes.

[0005] Furthermore, obtaining the input features of the signal after electrical stimulation includes: constructing a data group, that is, taking the subject's seizure situation as the first data group and any data feature of the signal as the second data group; calculating the correlation coefficient between the first data group and the second data group; sorting the correlation coefficients of each data feature to select data features related to the seizure situation as the input features of the classifier.

[0006] Furthermore, obtaining the proportion of stimulation channels in which the energy of the seizure changes before and after electrical stimulation includes calculating the mean and variance of the energy characteristics of the seizure period before and after electrical stimulation to obtain the proportion of stimulation channels in which the seizure energy increases or decreases, that is, for signals that do not contain electrical stimulation: directly calculate the mean and variance of the energy characteristics of the seizure period before and after electrical stimulation; for signals that contain electrical stimulation: calculate the mean and variance of the energy characteristics of the seizure period before and after electrical stimulation after processing the stimulation artifacts.

[0007] Furthermore, the processing of stimulation artifacts includes: identifying electrical stimulation, including: differentiating the signal, identifying the position where the differential value exceeds the threshold as an abnormal point, classifying the abnormal points with adjacent intervals less than the threshold as the same electrical stimulation, and removing all data from the electrical stimulation period; removing the baseline, including: dividing the original data of the attack period into multiple segments, using a monotonic fitting method to fit the baseline to each segment of data, and subtracting the fitted baseline from the original data.

[0008] Furthermore, obtaining the situation where the stimulation channels cover the strongly connected nodes includes constructing a network: taking the stimulation channels as network nodes, the connections between the nodes as the edges of the network, and calculating the phase lag index between the stimulation channels as the connection strength of the edge during the attack period; identifying the strongly connected nodes: for each attack period, retaining the edges whose connection strength exceeds the threshold, marking the nodes connected by the edges, and counting the number of times each node appears in all attack periods; marking the nodes whose number of appearances exceeds the threshold as strongly connected nodes; and judging whether the stimulation channels cover the strongly connected nodes.

[0009] Furthermore, it also includes: a processor; the processor trains a classifier based on the input features and the classification results of the effectiveness of the electrical stimulation to pre-adjust the stimulation mode of the electrical stimulation.

[0010] In a second aspect, the present invention provides a neuromodulatory device comprising: electrodes, a stimulation module and the electrical stimulation system; wherein the processor controls the stimulation module to call a corresponding stimulation mode to implement electrical stimulation through the electrodes.

[0011] In a third aspect, the present invention provides an epilepsy treatment device, comprising: a first device, located outside the body, comprising the electrical stimulation system, for adjusting the stimulation pattern of electrical stimulation; a second device, located outside the body, containing a power supply module; a third device, located inside the body, comprising a stimulation module and electrodes; wherein the third device is wirelessly coupled to the second device to obtain electrical energy through the power supply module; and the third device wirelessly communicates with the first device to enable the stimulation module to obtain the stimulation pattern.

[0012] The beneficial effect of the present invention is that the electrical stimulation system, neural regulation device, and epilepsy treatment device of the present invention set the input end of the classifier to the input feature of the signal after electrical stimulation, and set its output end to the classification result of the effectiveness of electrical stimulation. The correlation coefficient between the data features and the effectiveness of electrical stimulation is used to obtain the proportion of stimulation channels with changes in seizure energy before and after electrical stimulation and\or the situation where the stimulation channels cover strong connection nodes, and use this as input features to train the classifier, so that the stimulation pattern and the stimulation effect are associated and bound according to the data features of the EEG signal. The effectiveness of the stimulation pattern can be effectively predicted in vitro or before the device is implanted, and the stimulation pattern can be adjusted.

[0013] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purposes and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0014] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 This is the working flow diagram of the electrical stimulation system.

[0017] Figure 2 It is the parameter setting interface for training the classifier of the electrical stimulation system.

[0018] Figure 3 It is the parameter setting interface for online stimulation of the electrical stimulation system.

[0019] Figure 4 This is a workflow diagram of the neuromodulation device.

[0020] Figure 5 It is the parameter setting interface of the closed-loop neural regulation device.

[0021] Figure 6 It is the EEG signal waveform of online electrical stimulation of closed-loop neural control equipment.

[0022] Figure 7 It is the parameter setting interface of the open-loop neural control device.

[0023] Figure 8 This is the principle block diagram of epilepsy treatment equipment.

[0024] Figure 9 Results of the effectiveness of the cross-validated stimulation patterns are shown. DETAILED DESCRIPTION

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0026] See Figure 1-Figure 3 , this embodiment provides an electrical stimulation system, including: a processor, a classifier connected to the processor, a human-computer interaction machine, etc. Specifically, the processor trains the classifier based on the input features of the historical signal and the classification results of the effectiveness of the electrical stimulation to pre-adjust the stimulation mode of the electrical stimulation to achieve online detection and electrical stimulation. The input end of the classifier is the input feature of the signal after electrical stimulation, and the output end is the classification result of the effectiveness of the electrical stimulation; wherein the input feature includes the proportion of stimulation channels with energy changes before and after the electrical stimulation and\or the situation where the stimulation channels cover strongly connected nodes.

[0027] Optionally, the electrical stimulation system is provided with PC hardware and PC software, and each algorithm step is run by a processor. The human-computer interaction machine is provided with an operation interface and a display, which can not only set operating parameters but also present waveform diagrams of electrophysiological signals before and after electrical stimulation.

[0028] Taking EEG signals as an example, the application of the electrical stimulation system in evaluating the effect of epilepsy stimulation is described, which specifically includes the following steps.

[0029] Step S1, obtaining input features of the signal after electrical stimulation.

[0030] (1) Obtain the proportion of stimulation channels with changes in seizure energy before and after electrical stimulation. Compare the seizure energy of each channel before and after stimulation to indicate the range of channels affected by the stimulation, including: calculating the mean and variance of the energy characteristics of the seizure period before and after electrical stimulation to obtain the proportion of stimulation channels with increased or decreased seizure energy. Specifically, for signals that do not contain electrical stimulation: directly calculate the mean and variance of the energy characteristics of the seizure period before and after electrical stimulation; for signals that contain electrical stimulation: calculate the mean and variance of the energy characteristics of the seizure period before and after electrical stimulation after processing the stimulation artifacts.

[0031] Optionally, for signals that do not contain electrical stimulation: intercept the EEG during the attack period, perform 1-120 Hz bandpass filtering, calculate the line length and RMS of the EEG as energy features, calculate the energy mean and variance of the entire attack period, and use them as the energy evaluation of the channel during the attack period.

[0032] Optionally, for signals containing electrical stimulation: intercept the EEG during the seizure period, perform electrical stimulation identification, baseline removal, 1-120Hz bandpass filtering in sequence, and calculate the mean and variance of the EEG energy characteristics. Identifying electrical stimulation includes: first differentiating the EEG, identifying the positions where the differential value exceeds 1000 as abnormal points, and considering the abnormal points with adjacent intervals less than the threshold to belong to the same electrical stimulation, removing all data during the electrical stimulation period, and dividing the data of the entire seizure period into multiple segments. Removing the baseline includes: using a monotonic fitting method to complete the fitting of the artifact baseline in each segment of data, and subtracting the fitted baseline from the original data.

[0033] (2) Obtain the situation of strong connection nodes covered by the stimulation channels. Construct the network: take the stimulation channels as network nodes, and the connections between nodes as network edges. Calculate the phase lag index (PLI) between each stimulation channel as the connection strength of the edge during the attack period; Identify strong connection nodes: For each attack period, retain the edges with connection strength exceeding the threshold (generally set to 50% of the highest connection strength), mark the nodes connected by the edges, and count the number of times each node appears in all attack periods; mark the nodes with a number of appearances exceeding the threshold as strong connection nodes; Determine whether the stimulation channels cover strong connection nodes, especially whether the stimulation electrodes contain strong connection nodes in single-point or multi-point stimulation.

[0034] (3) Obtain the input features of the signal after electrical stimulation. Construct a data group, that is, take the subject's seizure situation as the first data group and any data feature of the signal as the second data group; calculate the correlation coefficient between the first data group and the second data group; sort the correlation coefficients of each data feature to select the data features related to the seizure situation as the input features of the classifier. Specifically, the seizure situation of the subject, such as the seizure frequency reduction rate and duration reduction rate, can be selected as positive evaluation indicators of the effectiveness of stimulation. Then, the data feature with the largest correlation coefficient with the seizure situation is selected as the input feature of the classifier, that is, the proportion of stimulation channels with seizure energy changes before and after electrical stimulation and\or the situation where the stimulation channels cover strong connection nodes. The model parameters between the stimulation mode, input features, and seizure situation (i.e., stimulation effectiveness) are obtained as stable indicators for evaluating the relationship between the stimulation mode and the stimulation effect and stored in the database, as shown in Table 1. The table shows the correlation coefficients between the data features of different EEG signals and the descriptive indicators of the treatment effect (seizure frequency reduction rate, duration reduction rate, stimulation effectiveness). The positive or negative value of the correlation coefficient indicates positive or negative correlation. The absolute value of the correlation coefficient indicates the strength of the correlation between the data feature and the treatment effect. The larger the absolute value of the correlation coefficient, the stronger the correlation between the corresponding data feature and the treatment effect, and the more accurately the data feature can be used to describe the treatment effect. The absolute values ​​of the correlation coefficients between "stimulating strongly connected PLI nodes" and "seizure frequency reduction rate" and "energy increase / reduction channel ratio" and "duration reduction rate" were large (>0.8), indicating that these two data features can represent the treatment effect from different perspectives.

[0035] Table 1 Stability indicators between input features and stimulation effects.

[0036] Data characteristics Seizure frequency reduction rate Duration reduction rate Stimulus effectiveness PLI pre-stimulus connection strength 0.22 0.35 0.29 Connection strength after PLI stimulation 0.11 0.31 0.12 Change rate of connection strength before and after PLI stimulation 0.22 0.32 0.24 PLI pre-stimulus global efficiency 0.32 0.27 0.18 Global efficiency after PLI stimulation 0.06 0.08 0.22 The change rate of global efficiency before and after PLI stimulation 0.15 -0.1 0.25 Interval discharge frequency reduction rate 0.71 0.44 0.56 … … … … Stimulate PLI strong connection nodes 0.91 -0.15 0.45 Multi-point stimulation of PLI strong connection nodes 0.28 -0.44 0.31 Mean connection strength after PLI stimulation 0.23 -0.31 -0.07 Energy increase channel ratio / reduction channel ratio -0.12 -0.81 0.31

[0037] Step S2: training the classifier.

[0038] like Figure 2 As shown, the processor trains a classifier based on historical data input features and classification results of electrical stimulation effectiveness to adjust or preset the stimulation mode of electrical stimulation. Optionally, the stimulation mode includes but is not limited to: at least one of stimulation current parameters, stimulation sequence, stimulation time, stimulation frequency (stimulation interval, number of stimulations, rest time, number of cycles), stimulation channel, and stimulation target.

[0039] (1) Reading training data. On the human-computer interaction interface, select the data path - select the result save path - select the file: start reading the file, read the next one, click the drop-down box - file name list: drop-down box display - waveform display.

[0040] (2) Waveform display. In the waveform browsing in the coordinate graph, select the browsing box to operate: drag to adjust the channel, drag to adjust the time, adjust the scale, etc.

[0041] (3) Trigger configuration. Delete or replace the original trigger; add a trigger; save the trigger as a file, read the trigger file, etc.

[0042] (4) Filter parameter configuration: such as low-pass, high-pass, notch parameters, etc.

[0043] (5) Preprocessing and feature extraction. Preprocessing: Windowed baseline removal and filtering; Feature extraction: Extract and save the features used in the model; Read saved files: Check the box that features have been saved.

[0044] (6) Seizure detection training and adjustment of model parameters. Automatic training: Automatic training according to the existing process, and plotting the results and parameters of the optimal channel; Manual parameter change: Manually add or subtract the small window threshold and large window threshold; Specify channel: Plot the results and model parameters of the specified channel.

[0045] (7) Preliminary detection training and adjustment of model parameters. Automatic training: Automatic training according to the existing process, and plotting the results and parameters of the optimal channel; Manual parameter change: Manually add or subtract the large window threshold; Specify channel: Plot the results and model parameters of the specified channel.

[0046] Step S3: online stimulation.

[0047] See Figure 3 The processor controls the stimulation module online to call the corresponding stimulation mode to implement electrical stimulation through the electrodes. Set the following detection parameters on the human-computer interaction interface:

[0048] Stimulation sequence: Generate stimulation sequence in automatic mode or manual mode, displaying all electrodes (such as pin 1 to pin 8) and channels on each electrode (such as C1 to C64); generate stimulation sequence based on electrodes and channels: single channel, ascending sequence for all channels, descending sequence for all channels.

[0049] Parameter configuration, adding new stimulation channels: setting and adding new electrode pairs used for stimulation; parameter adjustment: stimulation interval, stimulation duration, stimulation current intensity, stimulation pulse width, pulse interval, stimulation frequency, train stimulation cycle, pulse duration, stimulation intensity step; parameter distribution: distribution to single / all electrode pairs.

[0050] Stimulation sequence configuration, stimulation sequence: list all stimulation pairs; number of cycle stimulation: stimulate all stimulation pairs and set the number of cycles; generate trigger and display specific parameters; delete stimulation pair: delete selected, delete all; trigger stimulation: stimulate the next channel pair, automatically stimulate in sequence, stimulate all at the same time, stimulate a single channel pair.

[0051] Discharge parameter configuration, manual discharge: duration, discharge channel; automatic discharge: discharge all channels after each stimulation.

[0052] Device status: Stimulation box connection status.

[0053] Stimulus mark trigger, stimulation name: stimulation channel pair; stimulation status: stimulation start / stimulation end; time: the time of the current trigger; stimulation information: record stimulation parameters.

[0054] In some embodiments, a neuromodulation device is provided, comprising: electrodes, a stimulation module, and the electrical stimulation system; wherein the processor controls the stimulation module to call a corresponding stimulation mode to implement electrical stimulation through the electrodes. Figure 4 The neuroregulatory device is divided into open-loop type (direct stimulation without online detection) and closed-loop type (deciding whether to implement stimulation after online detection). It uses EEG characteristics to predict the patient's treatment effect, guide the electrode placement position of the neuroregulatory device, and can guide the initial stimulation parameters to quickly verify the short-term treatment effect.

[0055] (1) Closed-loop neural control equipment.

[0056] See Figure 5 The workflow of the closed-loop neuromodulation device includes: configuring sEEG device parameters, setting preprocessing parameters, seizure detection parameters, preictal detection parameters, protection parameters, configuring ECS ​​device parameters, selecting stimulation targets, and confirming and starting the process. The detailed operation interface of the human-computer interface is shown below.

[0057] Basic configuration: whether to save data, save path, file name (not save); current number of channels of sEEG device, sampling rate (134, 2000); detection parameters: detection channel number, lead mode (81, bipolar lead); whether to output stimulation command: only collect (do not send stimulation command to the outside), stimulate (call stimulation mode).

[0058] Preprocessing parameter configuration, baseline drift: subtract the starting 0.1s mean value (0.1) from each segment of data; low-pass, high-pass, notch: filter coefficients a and b.

[0059] Seizure detection classifier parameter configuration: minimum window thresholds 1 and 2: 0.1s window energy during seizures and spikes (9000, 4750); window thresholds during seizures and spikes: number of consecutive abnormal windows during seizures and spikes (9, 8).

[0060] Parameter configuration of the pre-ictal detection classifier, parameter 1, parameter 2: pre-ictal detection classifier parameter vector, such as w and c of LDA; continuous window number threshold: in the pre-ictal period, the number of continuous abnormal windows (6).

[0061] Stimulation protection parameters, stimulation interval, number of stimulations: the output stimulation cannot exceed 10 times in 300s (300, 10).

[0062] Acquisition and stimulation device parameters: sEEG device: device IP and communication port (127.0.0.1, 8000); ECS device: device IP and communication port (127.0.0.1, 9000).

[0063] Select stimulation target, stimulate onset, stimulate pre-onset, stimulate spike wave: detect and stimulate onset and pre-onset.

[0064] Electrical stimulation: detect epileptic seizures online and automatically apply electrical stimulation until the seizure stops. Figure 6 Shown are the closed-loop stimulation results for the patient during the first and second episodes.

[0065] (2) Open-loop neural control equipment.

[0066] See Figure 7 The workflow of the open-loop neural control device includes: configuring sEEG device parameters, setting preprocessing parameters, protection parameters, open-loop parameters, and configuring ECS ​​device parameters. Figure 8 As shown, the following parameters can be set on the human-computer interaction interface.

[0067] Basic configuration: whether to save data, save path, file name (not save); current number of channels of sEEG device, sampling rate (115, 2000).

[0068] Preprocessing parameter configuration, baseline drift: subtract the starting 0.1s mean value (0.1) from each segment of data; low-pass, high-pass, notch: filter coefficients a and b.

[0069] Seizure detection classifier parameter configuration: minimum window threshold 1, 2; seizure and spike window threshold.

[0070] Preictal detection classifier parameter configuration: parameters 1 and 2; threshold of the number of consecutive windows.

[0071] Stimulation protection parameters, stimulation interval, number of stimulations: the output stimulation cannot exceed 10 times in 300s (300, 10).

[0072] Open-loop parameter configuration, such as stimulation interval, number of stimulations, rest time, and number of cycles: stimulation 1s, rest 9s, stimulation 6 times, rest 300s, and cycle 100 times (9, 6, 300, 100).

[0073] Acquisition and stimulation device parameters: sEEG device: device IP and communication port (127.0.0.1, 8000); ECS device: device IP and communication port (127.0.0.1, 9000).

[0074] Select stimulation target: Stimulate onset, Stimulate pre-onset, Stimulate spikes: detect and stimulate onset and pre-onset.

[0075] See Figure 8 In some embodiments, an epilepsy treatment device is provided, comprising: a first device, located outside the body, comprising the electrical stimulation system, for adjusting the stimulation pattern of electrical stimulation; a second device, located outside the body, comprising an energy supply module; a third device, located inside the body, comprising a stimulation module and electrodes; wherein the third device is wirelessly coupled to the second device to obtain electrical energy through the energy supply module; and the third device wirelessly communicates with the first device to enable the stimulation module to obtain the stimulation pattern.

[0076] Specifically, when the stimulation mode needs to be adjusted, the third device can be wirelessly communicated with the first device to adjust the stimulation mode of electrical stimulation using the first device; thereafter, the third device is disconnected from the first device, and electrical stimulation is implemented using the adjusted stimulation mode. In other words, the epilepsy treatment device in this case forwards, processes, and classifies EEG signals in real time, and controls the application of electrical stimulation based on the results. Before the third device is formally implanted, it can use historical data to verify the various algorithms in the third device in vitro, use EEG features to bind the stimulation mode, and then confirm the effect of neuromodulation therapy. Before and after the formal implantation of the third device, the third device is wirelessly coupled to the second device to obtain electrical energy through the energy supply module, and then the preset stimulation mode can be used to achieve online detection and stimulation, without the need for extensive machine training after implantation.

[0077] Comparison of test results.

[0078] Using multiple epilepsy patients (01-10) as examples, we used the proportion of stimulation channels that showed changes in seizure energy before and after electrical stimulation and / or the coverage of strongly connected nodes by stimulation channels as classifier input features, and the effectiveness of long-term neuromodulation as the classifier output label. A K-nearest neighbor (KNN) classifier was trained. Feature 1 (the proportion of channels with increased seizure energy / the proportion of channels with decreased seizure energy) and feature 2 (the coverage of strongly connected nodes by stimulation channels, i.e., whether key nodes were stimulated) served as the second data set, while the confusion matrix for predicting treatment effectiveness (reduction in seizure duration or frequency) served as the first data set, as shown in Table 2.

[0079] Table 2 Elements of the confusion matrix.

[0080] patient 01 02 03 04 05 06 07 08 09 10 Feature 1 1 1 0 1 1 1 1 1 1 1 Feature 2 0.79 0.58 0.37 0.70 0.99 0.40 0.47 0.11 0.99 0.86 Effectiveness 1 0 0 1 0 1 1 1 0 1

[0081] like Figure 9 As shown in Figure 3, the confusion matrix of the recognition results of the constructed simple classification model is used to describe the classification accuracy of the model to cross-validate the effectiveness results of the stimulation pattern. Figure 9 Middle: Patients with no treatment effect are labeled 0, and patients with treatment effect are labeled 1. The model accurately identifies patients with treatment effect with 100% accuracy (6 / 6); it accurately classifies patients with no treatment effect with 50% accuracy (2 / 4), resulting in an overall accuracy of 80% (8 / 10). This indicates that the classifier constructed using these two input features has some classification capability. With further data collection, the accuracy can be further improved. This shows that:

[0082] (1) Based on the short-term treatment effect during sEEG, a classifier can be used to effectively predict whether the patient is suitable for the stimulation mode (i.e., treatment plan).

[0083] (2) The therapeutic effect can be improved by stimulating key points.

[0084] (3) By checking whether energy reduction occurs after short-term treatment, it can be used to adjust the stimulation location / stimulation parameters.

[0085] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a portion of code, and the module, program segment or a portion of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or can be implemented using a combination of dedicated hardware and computer instructions.

[0086] If the function or model is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0087] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is merely a logical functional division, and actual implementation may have other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not implemented.

[0088] 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 scope of this invention. The technical scope of this invention is not limited to the contents of the specification and must be determined according to the scope of the claims.

Claims

1. An electrical stimulation system comprising: A classifier; characterized in that The input end of the classifier is the input feature of the signal after electrical stimulation, and the output end is the classification result of the effectiveness of electrical stimulation; The input features include the proportion of stimulation channels with energy changes before and after electrical stimulation and / or the situation where the stimulation channels cover strongly connected nodes; The input features of the signal after electrical stimulation include: Constructing a data set, that is, taking the seizure condition of the subject as the first data set and taking any data feature of the signal as the second data set; Calculating a correlation coefficient between the first data set and the second data set; Sort the correlation coefficients of each data feature to select the data features related to the attack situation as the input features of the classifier; The seizure status of the subject is configured as seizure frequency reduction rate and / or duration reduction rate; The data features are configured to stimulate strong connection nodes and / or energy increase channel ratio and / or energy decrease channel ratio; The correlation coefficient is configured to include at least one of a correlation coefficient between stimulation of strongly connected nodes and seizure frequency reduction rate, a correlation coefficient between energy increase channel ratio and duration reduction rate, and a correlation coefficient between energy decrease channel ratio and duration reduction rate; The situations where the stimulation channel covers the strongly connected nodes include: Construct a network: The stimulation channels are regarded as network nodes, and the connections between nodes are network edges. The phase lag index between each stimulation channel is calculated as the connection strength of the edge during the seizure period. Identify strongly connected nodes: For each attack period, retain the edges whose connection strength exceeds the threshold, mark the nodes connected by the edges, and count the number of times each node appears in all attack periods; mark the nodes whose number of appearances exceeds the threshold as strongly connected nodes; Determine whether the stimulation channel covers a strongly connected node.

2. The electrical stimulation system according to claim 1, wherein The proportion of stimulation channels for obtaining the change in seizure energy before and after electrical stimulation includes: The mean and variance of the energy characteristics of the seizure period before and after electrical stimulation were calculated to obtain the proportion of stimulation channels with increased or decreased seizure energy, i.e. For signals that do not contain electrical stimulation: directly calculate the mean and variance of the energy characteristics of the attack period before and after electrical stimulation; For signals containing electrical stimulation: after processing the stimulation artifacts, calculate the mean and variance of the energy characteristics of the attack period before and after the electrical stimulation.

3. The electrical stimulation system according to claim 2, wherein: The processing of stimulus artifacts includes: Identifying electrical stimulation, including: performing differential analysis on the signal, identifying the position where the differential value exceeds a threshold as an abnormal point, classifying the abnormal points whose adjacent intervals are less than the threshold as the same electrical stimulation, and removing all data from the electrical stimulation period; The baseline removal includes: dividing the original data of the attack period into multiple segments, fitting the baseline of each segment data using a monotone fitting method, and subtracting the fitted baseline from the original data.

4. The electrical stimulation system according to claim 1, wherein Also includes: processor; The processor trains a classifier based on the input features and the classification result of the electrical stimulation effectiveness to adjust a stimulation pattern of the electrical stimulation.

5. A neural regulation device, characterized in that: include: Electrodes, stimulation modules and an electrical stimulation system as claimed in claim 4; in The processor controls the stimulation module to call a corresponding stimulation mode to implement electrical stimulation through the electrodes.

6. An epilepsy treatment device, characterized in that: include: A first device, located outside the body, comprising the electrical stimulation system according to any one of claims 1 to 4, for pre-adjusting a stimulation pattern of electrical stimulation; The second device, located outside the body, contains the energy supply module; a third device, located inside the body, comprising a stimulation module and electrodes; in The third device is wirelessly coupled to the second device to obtain electrical energy through the energy supply module; The third device communicates wirelessly with the first device so that the stimulation module acquires the stimulation pattern.

Images