Method for positioning epileptic focus by SEEG electrode and processing device

Abstract

The invention provides a SEEG electrode positioning epileptic focus method and a processing device, and the method comprises the steps: constructing an individualized digital twin brain model based on multi-modal imaging data and physiological signal data; based on the digital twin brain model, obtaining a possible epilepsy region; obtaining an optimal SEEG electrode distribution scheme based on the possible epilepsy region; based on the optimal SEEG electrode distribution scheme, the epilepsy focus is positioned; according to the method, the possible epilepsy region is positioned by constructing the digital twin brain model, the association between the brain structure abnormality and the physiological signal abnormality is comprehensively considered, and the optimal SEEG electrode distribution scheme is determined according to the possible epilepsy region, so that the SEEG electrodes cover the suspicious region more accurately, the number of the SEEG electrodes is reduced, the cost burden of a patient is reduced, and the accuracy of the SEEG electrode distribution scheme is improved. And the monitoring effectiveness and the data processing efficiency are improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent biomedical signal processing, and in particular to a method and a processing device for locating epileptogenic foci using a SEEG electrode. Background Art

[0002] Surgical treatment of drug-refractory epilepsy is a functional neurosurgery technique that has developed rapidly in recent years. One of its core technologies is to accurately locate the epileptogenic focus. Some epilepsy patients can locate the epileptogenic focus based on the typical clinical manifestations of epileptic seizures, head CT, MRI, PET and other imaging data, and non-invasive examinations such as electroencephalograms, but some patients need invasive electroencephalograms to locate the epileptogenic focus, including two invasive examination methods: cortical electrodes and deep electrodes. Cortical electrode monitoring requires a large craniotomy with a bone flap, and the cortical electrodes are placed on the surface of the brain tissue. This is traumatic, bleeding, and prone to intracranial infection. In addition, a second craniotomy is required to remove the cortical electrodes, and its clinical application is gradually decreasing.

[0003] Stereo electroencephalogram (SEEG) technology uses stereotactic technology or neurosurgery robot-assisted technology to accurately place deep electrodes in the brain area related to the origin or propagation path of epileptic seizures. It can not only collect EEG signals and map brain functional areas, but also use radiofrequency thermal coagulation to destroy epileptic foci, achieving the dual purposes of diagnosis and treatment. It represents the highest level of epileptic foci localization at present, with the advantages of minimally invasive, light damage, and the ability to provide deep EEG information. However, intracranial electrodes are expensive, so how to locate epileptic foci with as few intracranial electrodes as possible has become a key issue that needs to be solved. Summary of the invention

[0004] In order to overcome the problem of how to solve the problem of epileptogenic focus localization with as few intracranial electrodes as possible in the prior art, the present application provides a method and processing device for locating epileptogenic focus with SEEG electrodes, which adopts the following technical solutions:

[0005] In a first aspect, the present application provides a method for locating an epileptogenic focus using a SEEG electrode, comprising:

[0006] Build an individualized digital twin brain model based on multimodal imaging data and physiological signal data;

[0007] Based on the digital twin brain model, obtaining possible epileptogenic areas;

[0008] Based on the possible epileptogenic zone, obtaining an optimal SEEG electrode distribution plan;

[0009] Based on the optimal SEEG electrode distribution plan, the epileptogenic focus was located.

[0010] Furthermore, the construction of an individualized digital twin brain model based on multimodal imaging data and physiological signal data includes:

[0011] Standardizing and registering the multimodal imaging data, aligning the data of different modalities into the same coordinate system, and obtaining preprocessed data;

[0012] Segmenting the brain image of the preprocessed data into different anatomical structures, and reconstructing the segmented brain structure data into a three-dimensional brain structure model;

[0013] Based on the fMRI data in the preprocessed data, identifying the functional connections between different brain regions and obtaining a brain functional network diagram;

[0014] Based on source reconstruction technology, EEG electrophysiological signals are mapped to a three-dimensional brain structure model to identify the electrical activities of different brain regions;

[0015] The three-dimensional brain structure model, brain functional network diagram and electrical activities of different brain regions are integrated to obtain a multi-dimensional digital twin brain model.

[0016] Furthermore, obtaining a possible epileptogenic area based on the digital twin brain model includes:

[0017] Perform time-frequency analysis on the EEG signal in each time window to obtain the spectrum characteristics of different time periods and identify the frequency components related to epileptic seizures;

[0018] Construct a functional connection network between different brain regions, take each brain region as a node, calculate the phase locking value between different brain regions, use the phase locking value between different brain regions as the weight of the edge, obtain the sub-network with abnormally high connection strength (high phase locking value), and determine the potential epileptogenic area;

[0019] The frequency components related to epileptic seizures and potential epileptogenic areas are input into the machine learning model for training, and the labeled epilepsy data is used for classification to finally predict the possible epileptogenic areas.

[0020] The final predicted possible epileptogenic zone location is mapped to the digital twin brain model, and based on the anatomical structure and functional connectivity, the marked area in the digital twin brain model is dynamically observed to determine the epileptogenic zone in the digital twin brain model.

[0021] Furthermore, obtaining an optimal SEEG electrode distribution scheme based on the possible epileptogenic zone includes:

[0022] Based on the epileptogenic zone, determine the target brain area that needs to be monitored;

[0023] Generate multiple possible distribution paths in the digital twin brain model based on preset constraints and target brain regions;

[0024] Find the optimal SEEG electrode distribution combination to minimize the number of SEEG electrodes covering the epileptogenic zone.

[0025] Furthermore, the method of finding the optimal SEEG electrode distribution combination to minimize the number of SEEG electrodes covering the epileptogenic zone includes:

[0026] Based on multiple possible distribution paths, a set of initial electrode distribution combinations is randomly generated as a population. Each distribution combination is regarded as an individual, and the objective function value of each individual is calculated. The higher the objective function, the better the individual.

[0027] According to the objective function value of the individual, a selection operator is used to select some individuals as parents. Individuals with high objective function values ​​are more likely to be selected.

[0028] Perform crossover operations on parent individuals, such as single-point crossover or multi-point crossover, to generate offspring individuals; the crossover operation simulates the gene recombination process to generate new distribution combination possibilities;

[0029] Replace some of the parent individuals with the offspring individuals to form a new population, and calculate the objective function value of each individual in the new population;

[0030] When the preset maximum number of iterations is reached or the objective function value of the best individual in the population does not change significantly in multiple iterations, the iteration is stopped and the optimal electrode distribution combination is output.

[0031] Furthermore, the objective function min(N) in the optimal electrode distribution combination needs to satisfy C ≥ C min and S ≥ S min , where N is the number of electrodes, min(N) is the minimum number of electrodes, C is the coverage index of the electrode combination, S is the safety factor, and C min is a pre-set minimum coverage, S min is a pre-set minimum safety factor.

[0032] Furthermore, the locating of the epileptogenic focus based on the optimal SEEG electrode distribution scheme includes:

[0033] The space of the digital twin brain model is divided into a number of voxels, the voxels are numbered, and the spatial position of each voxel is recorded; the voxels constitute a potential source space, and each voxel is regarded as a possible source of electrical activity; each voxel is assigned a conductivity, and the conductivity is obtained based on the imaging data of the individual;

[0034] Based on the geometric shape of the brain, tissue conductivity, and the positional relationship between voxels and SEEG electrodes, the potential generated by each voxel at the electrode position is obtained to obtain a potential distribution matrix, which represents the potential contribution generated by each voxel source at all electrode positions; each row of the positioning distribution matrix corresponds to a SEEG electrode, and each column corresponds to a voxel source in the brain space;

[0035] Based on the optimal SEEG electrode distribution combination, the EEG signals of the patient during the attack and the inter-attack period are collected, and the EEG signals are pre-processed, including filtering the EEG signals and removing artifacts to obtain the pre-processed EEG signals;

[0036] Perform feature extraction on the preprocessed EEG signal to obtain the time domain features, frequency domain features and time-frequency features of the EEG signal;

[0037] A multi-constrained linear equation system is constructed based on time domain features, frequency domain features and time-frequency features, and the regularization strategy is adjusted for the equation system. The equation system constructed based on time domain, frequency domain and time-frequency features and the equation system adjusted by the regularization strategy are jointly solved to obtain the estimated value of the voxel source intensity vector x. The location of the epileptic focus is determined based on the estimated value, and the voxel with the maximum source intensity or the voxel area is determined as the possible location of the epileptic focus.

[0038] In a second aspect, the present application also provides a SEEG electrode localization epileptogenic focus treatment device, comprising:

[0039] A digital twin brain model construction module, which is used to build an individualized digital twin brain model based on multimodal imaging data and physiological signal data;

[0040] A possible epileptogenic zone acquisition module, used to acquire the possible epileptogenic zone based on the digital twin brain model;

[0041] An optimal SEEG electrode distribution scheme acquisition module, used for acquiring an optimal SEEG electrode distribution scheme based on the possible epileptogenic zone;

[0042] The epileptogenic focus localization module is used to locate the epileptogenic focus based on the optimal SEEG electrode distribution plan.

[0043] In a third aspect, the present application provides an electronic device, including:

[0044] One or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions, which, when executed by the device, cause the device to perform the method as described in the first aspect.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer-readable storage medium is run on a computer, the computer executes the method described in the first aspect.

[0046] In a fifth aspect, the present application provides a computer program, which, when executed by a computer, is used to execute the method described in the first aspect.

[0047] In one possible design, the program in the fifth aspect may be stored in whole or in part on a storage medium packaged together with the processor, or may be stored in whole or in part on a memory not packaged together with the processor.

[0048] This application has the following beneficial effects:

[0049] 1. This application constructs an individualized digital twin brain model based on multimodal imaging data and physiological signal data, so that in the process of identifying possible epileptogenic areas and locating epileptogenic foci, it does not rely solely on imaging data and EEG data. By building a digital twin brain model to combine physiological signal data with imaging data, the regional and dynamic changes of brain function are further clarified. When abnormal rhythm of EEG signals is observed in a certain brain area, the potential connection between abnormal electrical activity and brain structure is analyzed in combination with the imaging characteristics of the area.

[0050] 2. The present application obtains the possible epileptogenic zone based on the digital twin brain model, and obtains the optimal SEEG electrode distribution scheme based on the possible epileptogenic zone. In the process of obtaining the possible epileptogenic zone, the present application comprehensively considers the complex relationship between brain structure abnormalities and physiological signal abnormalities, avoids excessive attention to irrelevant areas and unnecessary examinations, and ensures that the electrodes collect signals in the most critical areas by determining the optimal SEEG electrode distribution scheme based on the possible epileptogenic zone. At the same time, when obtaining the optimal SEEG electrode distribution scheme, the minimization of the number of electrodes, coverage, and safety factor are used as constraints, so that the SEEG electrode distribution scheme satisfies the minimum number of electrodes, ensures the coverage and safety of the epileptogenic zone, reduces the cost burden on patients, and improves the effectiveness of monitoring and the efficiency of data processing.

[0051] 3. The present application constructs a multi-constrained linear equation group based on time domain features, frequency domain features and time-frequency features, and adjusts the regularization strategy of the equation group, jointly solves the equation group constructed based on time domain, frequency domain and time-frequency features and the equation group adjusted by the regularization strategy, obtains the estimated value of the voxel source intensity vector x, determines the location of the epileptic focus based on the estimated value, and determines the voxel source with the maximum intensity or the voxel area as the possible location of the epileptic focus; the electrode channel corresponding to the onset time of the seizure in the time domain features is often closely related to the epileptic focus. By constructing an equation group containing time domain features, these brain areas related to the starting signal can be given priority; in the frequency domain features, for the voxel sources corresponding to the energy abnormality areas in the frequency domain, corresponding constraints are set in the equation group, so that the final voxel source intensity estimation is more consistent with the frequency characteristics of the actual EEG signal, thereby improving the positioning accuracy; according to the propagation path and frequency change pattern of the electrical activity in the time-frequency diagram, time-varying constraints are set in the equation group, so as to more accurately track the position changes of the epileptic focus and the dynamic information of the epileptic network. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is an exemplary system architecture diagram to which the embodiments of the present application can be applied;

[0053] Figure 2 This is a flow chart of a method for locating epileptogenic focus using SEEG electrodes according to an embodiment of the present application;

[0054] Figure 3 A flowchart for constructing a digital twin brain model according to an embodiment of the present application;

[0055] Figure 4 A flowchart for obtaining a possible epileptogenic zone according to an embodiment of the present application;

[0056] Figure 5 A flowchart for obtaining the optimal SEEG electrode distribution scheme for the embodiment of the present application;

[0057] Figure 6 A flowchart for obtaining the optimal electrode distribution combination for an embodiment of the present application;

[0058] Figure 7 This is a flow chart of epileptogenic focus localization according to an embodiment of the present application;

[0059] Figure 8 A flow chart of a processing device according to an embodiment of the present application;

[0060] Fig. 9 It is a schematic diagram of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0062] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0063] 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 accompanying drawings.

[0064] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0065] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0066] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III, Moving Picture Experts Group Audio Layer 3), MP4 (Moving Picture Experts Group Audio Layer IV, Moving Picture Experts Group Audio Layer 4) players, laptop computers and desktop computers, etc.

[0067] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0068] It should be noted that the SEEG electrode localization epileptic focus method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the SEEG electrode localization epileptic focus processing device is generally arranged in the server / terminal device.

[0069] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0070] Continue to refer Figure 2 , the figure shows a flow chart of a method for locating epileptogenic focus using SEEG electrodes of the present application, the method comprising the following steps:

[0071] Step 201: construct an individualized digital twin brain model based on multimodal imaging data and physiological signal data.

[0072] In a possible implementation, imaging data acquisition includes at least one of magnetic resonance imaging, functional magnetic resonance imaging, computed tomography, diffusion tensor imaging, etc. In actual scenarios, imaging data is acquired according to actual needs.

[0073] Explanation: Magnetic resonance imaging (MRI) can obtain structural brain images, functional magnetic resonance imaging (fMRI) can obtain brain functional activity data, computed tomography (CT) can obtain anatomical structure data, and diffusion tensor imaging (DTI) can obtain structural connection information of brain white matter fiber bundles;

[0074] In a possible implementation, the physiological signal data collection includes at least one of electroencephalogram (EEG), magnetoencephalogram (MEG), functional near-infrared spectroscopy (fNIRS), etc. In actual scenarios, the physiological signal data is collected according to actual needs.

[0075] Explanation: Electroencephalogram (MRI) records the electrophysiological activity of the brain and can be collected in different states (such as wakefulness, sleep, interictal period and ictal period) to obtain information on the characteristics of epileptic seizures and the functional state of the brain; magnetoencephalography (MEG) has a high temporal resolution and can be used to detect changes in the brain's magnetic field. When combined with electroencephalography (EEG), it can improve the accuracy of localizing the source of epilepsy; functional near-infrared spectroscopy (fNIRS) can measure changes in blood oxygenation in local areas of the brain and provide information about brain functional activity.

[0076] In one possible implementation, a personalized digital twin brain model is constructed based on multimodal imaging data and physiological signal data. Please refer to Figure 3 , including:

[0077] Step 11, standardize and register the multimodal imaging data, align the data of different modalities into the same coordinate system, and obtain preprocessed data.

[0078] Step 12, segmenting the brain image of the preprocessed data into different anatomical structures, reconstructing the segmented brain structure data into a three-dimensional brain structure model, the three-dimensional brain structure model includes structures such as the brain surface, ventricles, and cerebral sulci, and performing surface smoothing on the three-dimensional brain structure model to improve visualization. The anatomical structure includes gray matter, white matter, ventricles, etc. of the brain.

[0079] Step 13, based on the fMRI data in the preprocessed data, identifying the functional connections between different brain regions and obtaining a brain functional network diagram.

[0080] Step 14, mapping the EEG electrophysiological signals into the three-dimensional brain structure model based on source reconstruction technology to identify the electrical activities of different brain regions.

[0081] Step 15: Fuse the three-dimensional brain structure model, the brain function network diagram, and the electrical activities of different brain regions to obtain a multi-dimensional digital twin brain model.

[0082] In a possible implementation, the multimodal imaging data and physiological signal data include data during normal times and during attacks.

[0083] Step 202, obtaining a possible epileptogenic area based on the digital twin brain model;

[0084] For steps to obtain possible epileptogenic zones based on the digital twin brain model, please refer to Figure 4 , the specific contents include:

[0085] Step 21, performing time-frequency analysis on the EEG signal in each time window, obtaining the spectrum characteristics of different time periods, and identifying the frequency components related to epileptic seizures.

[0086] In a possible implementation, short-time Fourier transform or wavelet transform may be used to perform time-frequency analysis on the EEG signal in each time window.

[0087] Explanation: Short-time Fourier transform is a method for analyzing signals in the time-frequency plane. It divides the signal into multiple small time windows, performs Fourier transform in each window, and obtains the energy distribution of the signal at different times and frequencies. For EEG signals, epileptic seizures are usually accompanied by abnormal energy changes in a specific frequency range (such as high-frequency oscillations, HFOs). For example, in temporal lobe epilepsy, high-frequency oscillations of 80-200Hz may occur. Through short-time Fourier transform, the EEG signal can be converted into a time-frequency diagram to observe the time and location of these abnormal frequencies.

[0088] Explanation: Wavelet transform can adaptively analyze the time-frequency characteristics of the signal. It uses wavelet functions of different scales to decompose the signal and uses different time resolutions for components of different frequencies. In epilepsy research, wavelet transform can better capture transient and non-stationary features in EEG signals, such as epileptic spikes and sharp waves. These abnormal waveforms will show obvious characteristics in the coefficient map after wavelet transform, which can help locate possible epileptogenic areas.

[0089] Step 22, constructing a functional connection network between different brain regions, taking each brain region as a node, calculating the phase locking value between different brain regions, taking the phase locking value between different brain regions as the weight of the edge, obtaining a subnetwork with abnormally high connection strength (high phase locking value), and determining the potential epileptogenic area;

[0090] Step 23, input the frequency components related to epileptic seizures and potential epileptogenic areas into the machine learning model for training, use the labeled epilepsy data for classification, and finally predict the possible epileptogenic areas;

[0091] Step 24, mapping the final predicted possible epileptogenic zone location into the digital twin brain model, dynamically observing the marked area in the digital twin brain model based on the anatomical structure and functional connectivity, and determining the epileptogenic zone in the digital twin brain model.

[0092] Step 203: Obtain an optimal SEEG electrode distribution plan based on the possible epileptogenic zone.

[0093] Explanation: The digital twin brain model is a three-dimensional space that contains structures such as the epileptogenic zone, important blood vessels, neural tissues, and brain functional areas. Each SEEG electrode needs to be implanted from the entry point on the surface of the brain to the vicinity of the target area. Based on the possible epileptogenic zone, the goal of obtaining the optimal SEEG electrode distribution plan is to find a set of electrode distribution combinations that cover the epileptogenic zone with the least number of SEEG electrodes while avoiding damage to important structures.

[0094] To obtain the optimal SEEG electrode distribution plan based on the possible epileptogenic zone, please refer to Figure 5 , the specific contents include:

[0095] Step 31, based on the epileptogenic zone, determining a target brain region that needs to be monitored intensively, wherein the target brain region that needs to be monitored intensively includes the size, shape, position of the epileptogenic zone and its relationship with surrounding important brain structures;

[0096] Step 32, generating multiple possible distribution paths in the digital twin brain model based on preset constraints and target brain regions;

[0097] The preset constraint condition is to avoid the minimum distance of important blood vessels and nerves, and ensure that the electrode does not enter the functional area or the low functional risk area.

[0098] Step 33, finding the optimal SEEG electrode distribution combination to minimize the number of SEEG electrodes covering the epileptogenic zone.

[0099] In one possible implementation, an objective function is defined to minimize the number of electrodes. If the number of electrodes is N, the objective function can be expressed as min(N); a coverage index C is defined to indicate the proportion of electrodes that effectively monitor the epileptogenic zone. Ideally, C should be close to 100%; a safety factor S is defined to measure the degree to which the path avoids important blood vessels, nerve tissues, and brain functional areas. The higher the value of S, the safer the path. For example, S = 1 means that all dangerous structures are completely avoided, and S = 0 means that the path directly passes through important structures. In summary, the objective function can be min(N), and at the same time, C ≥ C min (C min is a pre-set minimum coverage, such as 90%) and S ≥ S min (S min is a pre-set minimum safety factor, such as 80%).

[0100] In one possible embodiment, the optimal SEEG electrode distribution combination is found to minimize the number of SEEG electrodes covering the epileptogenic zone. Figure 6 , including:

[0101] Step 331, based on multiple possible distribution paths, randomly generate a set of initial electrode distribution combinations as a population, each distribution combination is regarded as an individual, and the objective function value of each individual is calculated, and the higher the objective function, the better the individual;

[0102] Step 332, according to the objective function value of the individual, use the selection operator to select some individuals as parents, and individuals with high objective function values ​​are more likely to be selected;

[0103] Step 333, performing a crossover operation on the parent individuals, such as a single-point crossover or a multi-point crossover, to generate offspring individuals; the crossover operation simulates the gene recombination process to generate new distribution combination possibilities;

[0104] Step 334, replace some of the parent individuals with the offspring individuals to form a new population, and calculate the objective function value of each individual in the new population;

[0105] Step 335: When the preset maximum number of iterations is reached or the objective function value of the best individual in the population does not change significantly in multiple iterations, the iteration is stopped and the optimal electrode distribution combination is output.

[0106] Step 204: localize the epileptogenic focus based on the optimal SEEG electrode distribution scheme.

[0107] For localization of epileptogenic foci based on the optimal SEEG electrode distribution scheme, please refer to Figure 7 , the specific contents include:

[0108] Step 41, dividing the space of the digital twin brain model into a number of voxels, numbering the voxels, and recording the spatial position of each voxel; the plurality of voxels constitute a potential source space, and each voxel is regarded as a possible source of electrical activity;

[0109] Step 42, assigning conductivity to each voxel, wherein the conductivity is obtained based on imaging data of the individual;

[0110] Step 43, based on the geometric shape of the brain, the conductivity of the tissue, and the positional relationship between the voxel and the SEEG electrode, the potential generated by each voxel at the electrode position is obtained to obtain a potential distribution matrix, which represents the potential contribution generated by each voxel source at all electrode positions; each row of the positioning distribution matrix corresponds to a SEEG electrode, and each column corresponds to a voxel source in the brain space; assuming that there are M electrodes for recording EEG signals, then the matrix has M rows. Assuming that the brain is divided into N voxel sources, then the matrix has N columns. The potential distribution matrix is ​​a mapping relationship that describes the potential transfer from the source space (voxel source) inside the brain to the external measurement space (electrode position), reflecting how the conductive properties of brain tissue, the geometric positional relationship between the voxel source and the electrode, and the electrical activity characteristics of the voxel source jointly determine the potential that can be measured at the electrode. For example, if a voxel source is close to an electrode and the conductivity of the brain tissue between them is high, then in the potential distribution matrix, the element value of the column corresponding to the voxel source in the row where the electrode is located may be relatively large, indicating that the voxel source contributes more to the potential of the electrode.

[0111] Step 44, based on the optimal SEEG electrode distribution combination, collecting the EEG signals of the patient during the attack period and the inter-attack period, preprocessing the EEG signals, including filtering the EEG signals and removing artifacts, to obtain preprocessed EEG signals;

[0112] Step 45, extracting features from the preprocessed EEG signal to obtain time domain features, frequency domain features, and time-frequency features of the EEG signal;

[0113] Step 46, construct a multi-constrained linear equation group based on time domain features, frequency domain features and time-frequency features, and adjust the regularization strategy for the equation group, jointly solve the equation group constructed based on time domain, frequency domain and time-frequency features and the equation group adjusted by the regularization strategy, obtain the estimated value of the voxel source intensity vector x, determine the location of the epileptic focus based on the estimated value, and determine the voxel with the maximum source intensity or the voxel area as the possible location of the epileptic focus.

[0114] Construction of multi-constrained linear equations:

[0115] In a possible implementation, a time domain equation is constructed: based on the onset time of the time domain feature and the time and position of the abnormal waveform, a constraint condition is constructed to construct the time domain equation A. t x=y t , where A t is the potential distribution matrix constructed based on time domain characteristics, y t For the actual measured electrode potential vector part related to the time domain characteristics (for example, mainly the electrode potential at the onset of the seizure), the time domain equation mainly constrains the potential contribution of the voxel source at the onset of the seizure, so that the result of the inverse solution can preferentially match the time domain signal characteristics at the onset of the seizure; x is the voxel source intensity vector.

[0116] In one possible implementation, construct the frequency domain characteristic equation: Incorporate the frequency domain characteristics into the equation system, then add equation A f x=y f , where A f is the potential distribution matrix constructed based on frequency domain features (for example, a matrix adjusted according to the propagation characteristics of different frequency electrical activities in the brain), y f is the part of the actual measured electrode potential vector related to the frequency domain characteristics (such as the potential corresponding to the energy of a specific frequency band), and x is the voxel source intensity vector.

[0117] In a possible implementation, based on the dynamic change of time-frequency characteristics, an equation that changes with time is constructed. Let A tf (t) is the potential distribution matrix constructed in time based on time-frequency characteristics, y tf (t) is the actual measured electrode potential vector part related to time and time-frequency characteristics. Then for different time points t1, t2, ..., t n , construct the equation system A tf (t i )x=y tf (t i), i = 1, 2, ..., n. The dynamic equation constrains the potential contribution of the voxel source during the entire attack process, so that it can match the dynamic changes of the EEG signal in time and frequency.

[0118] Regularization strategy adjustment:

[0119] In a possible implementation, the regularization strategy is adjusted based on the time domain characteristics, and the voxel sources around the seizure onset electrode are given a lower weight in the regularization process due to their importance in locating the starting position of the epileptic focus. For example, in Tikhonov regularization, the regularization parameter is set to λ, and the weight of the regularization term for the voxel sources in the key suspected area is set to λ1 (λ>λ1), so that the voxel sources have greater flexibility to fit the time domain signal of the seizure onset during the inverse solution process.

[0120] In one possible implementation, the regularization parameter is adjusted based on the energy distribution in the frequency domain features; if the energy distribution of a certain frequency band (such as high-frequency oscillation) is relatively concentrated in a few regions, the regularization parameter is reduced for the voxel sources corresponding to these regions. For example, the regularization parameter λ is adjusted according to the proportion p of the high-frequency oscillation energy to the total energy, and the adjusted regularization parameter is set to λ2=λ(1-θp) (θ is the adjustment coefficient), so that the inverse solution can pay more attention to the adjustment of the voxel source intensity in these frequency domain energy abnormal regions.

[0121] In one possible implementation, a dynamic regularization strategy is adopted according to the dynamic characteristics of the time-frequency features. During the propagation of the electrical activity, the regularization parameter is dynamically changed according to the frequency change and the propagation direction. For example, at the beginning of the propagation of the electrical activity, in order to better capture the location of the source, a smaller regularization parameter is used for the voxel source around the starting area; as the electrical activity propagates, the regularization parameter is gradually increased to stabilize the solution process and avoid overfitting.

[0122] In a possible implementation, the equations constructed based on the time domain, frequency domain and time-frequency features and the equations adjusted by the regularization strategy are jointly solved, which is specifically manifested as follows:

[0123] The equations constructed based on time domain, frequency domain and time-frequency features are combined into a large linear equation system. Assume that the equation system based on time domain features is A t x=y t , the frequency domain characteristic equation group is A f x=y f , the equation group of time-frequency characteristics is A tf (t i )x=y tf (t i )(i=1,2,...,n), they can be combined into a new system of equations Ax=y; where A is the equation of At , A f and A tf (t i ) is a matrix formed by concatenating rows, y is a matrix composed of y t ,y f and tf (t i ) is a vector concatenated by rows, x is a voxel source intensity vector, and its dimension depends on the number of voxel source divisions in the brain model.

[0124] Taking Tikhonov regularization as an example, construct the objective function J(x) = ||Ax-y|| 2 +λ||x|| 2 , λ is the regularization parameter. The first term of the objective function is to minimize the difference between the calculated electrode potential (obtained by forward calculation based on the voxel source intensity x) and the actual measured electrode potential y. The second term is the regularization constraint on the voxel source intensity vector x to prevent the elements of x from being too large or changing too drastically.

[0125] Perform singular value decomposition on matrix A: A=UΣV T , where U is an m×m orthogonal matrix (m is the number of rows of A), whose column vectors are called left singular vectors; Σ is an m×n diagonal matrix (n is the number of columns of A), the elements on the diagonal are called singular values, and are arranged in descending order; V is an n×n orthogonal matrix, whose column vectors are called right singular vectors; for any matrix A, A T A and AA T are symmetric semi-positive matrices whose eigenvalues ​​are the squares of the singular values ​​in Σ, and whose corresponding eigenvectors are the column vectors of V and U respectively.

[0126] Substitute x=Vz into the objective function J(x), and we get J(z)=||A(Vz)-y|| 2 +λ||Vz|| 2 , because A=UΣV T , we get J(z)=||UΣz-y|| 2 +λ||Vz|| 2 , because U is an orthogonal matrix, so ||UΣz-y|| 2 =||Σz-U T y|| 2 , and ||Vz|| 2 =z T z (because V is an orthogonal matrix), the objective function becomes J(z) = ||Σz-U T y|| 2 +λz T z, expand ||Σz-U T y|| 2Get (Σz-U T y) T (Σz-U T y)=z T Σ 2 z-2z T ΣU T y+y T UU T y, since U is an orthogonal matrix UU T =I, so y T UU T y=y T y, the objective function further becomes J(z)=z T Σ 2 z-2z T ΣU T y+y T y+λz T z, take the derivative of J(z) with respect to z and set the reciprocal to 0, and we get: After sorting, we get (Σ 2 +λI)z=ΣU T y, let (Σ 2 +λI)=diag(σ1 2 +λ,σ2 2 +λ,...,σ n 2 +λ)ΣU T y=[v1,v2,...,v n ] T Then the element z of z i Available through After solving and obtaining z, the voxel source intensity vector x is solved by x=Vz.

[0127] Explanation: The initial guess of voxel source intensity is constructed through the time domain characteristics, frequency domain characteristics and time-frequency characteristics of the EEG signal, specifically:

[0128] Based on the seizure onset time and electrode channel information in the time domain features, the brain region corresponding to the electrode channel where the seizure onset signal first appears is taken as the key suspected epileptogenic focus area; in the initial stage of the inverse solution, a higher initial source intensity is given to the voxel source near this area. For example, assume that the initial intensity of the voxel source within a radius of around the seizure onset electrode is set to x0, while the initial intensity of the voxel source in other areas is set to x1 (x0>x1).

[0129] According to the power spectral density (PSD) analysis in the frequency domain characteristics, an abnormal increase in high-frequency oscillation energy was found in the frequency domain signals of the electrode channels corresponding to the suspected epileptogenic focus area, which increased the initial intensity guess of the voxel source in the area; for example, based on the initial intensity x0, the high-frequency oscillation energy was increased proportionally according to the relative size of the high-frequency oscillation energy, such as the increase coefficient is k (k is related to the proportion of high-frequency oscillation energy).

[0130] The time-frequency characteristics are used to obtain the energy enhancement or change pattern of electrical activity at a specific frequency at the onset of the attack and the propagation of this change in different brain regions. When the electrical activity propagates from an initial suspected area to a specific direction, the initial intensity is dynamically adjusted according to the frequency change during the propagation process. For example, on the propagation path, as the frequency decays, the initial guess of the voxel source intensity decreases according to a certain decay function.

[0131] The goal of this application through inverse solution is to infer the actual intensity of each voxel source inside the brain based on the actual measured electrode potential (the patient's EEG signal recorded by the SEEG electrode). By combining the actual measured electrode potential with the potential distribution matrix obtained by forward calculation, mathematical methods (such as constructing a linear equation system and solving it) can be used to estimate the intensity of each voxel source, and then find the possible location of the epileptogenic focus (usually the voxel or voxel area with the largest source intensity). Map the voxel source intensity to the digital twin model of the brain to intuitively observe the location of the epileptogenic focus.

[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0133] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0134] Continue to refer Figure 8 The SEEG electrode positioning epileptogenic focus treatment device described in this embodiment includes:

[0135] A digital twin brain model construction module 801 is used to construct an individualized digital twin brain model based on multimodal imaging data and physiological signal data;

[0136] A possible epileptogenic zone acquisition module 802 is used to acquire a possible epileptogenic zone based on the digital twin brain model;

[0137] An optimal SEEG electrode distribution scheme acquisition module 803 is used to acquire an optimal SEEG electrode distribution scheme based on the possible epileptogenic zone;

[0138] The epileptogenic focus localization module 804 is used to localize the epileptogenic focus based on the optimal SEEG electrode distribution scheme.

[0139] To solve the above technical problems, the present application also provides a computer device. Fig. 9 , Fig. 9 This is a basic structural block diagram of the computer device in this embodiment.

[0140] The computer device 9 includes a memory 9a, a processor 9b, and a network interface 9c that are interconnected through a system bus. It should be noted that the figure only shows a computer device 9 with components 9a-9c, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0141] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0142] The memory 9a includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 9a can be an internal storage unit of the computer device 9, such as a hard disk or memory of the computer device 9. In other embodiments, the memory 9a can also be an external storage device of the computer device 9, such as a plug-in hard disk equipped on the computer device 9, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 9a can also include both the internal storage unit of the computer device 9 and its external storage device. In this embodiment, the memory 9a is generally used to store the operating system and various application software installed on the computer device 9, such as the program code of the SEEG electrode localization epileptogenic focus method, etc. In addition, the memory 9a can also be used to temporarily store various types of data that have been output or are to be output.

[0143] The processor 9b may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 9b is generally used to control the overall operation of the computer device 9. In this embodiment, the processor 9b is used to run the program code stored in the memory 9a or process data, such as running the program code of the SEEG electrode localization epileptogenic focus method.

[0144] The network interface 9c may include a wireless network interface or a wired network interface. The network interface 9c is generally used to establish a communication connection between the computer device 9 and other electronic devices.

[0145] The present application also provides another embodiment, namely, providing a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a program of a method for locating an epileptic focus with a SEEG electrode, and the program can be executed by at least one processor so that the at least one processor performs the steps of the method for locating an epileptic focus with a SEEG electrode as described above.

[0146] 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, 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, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0147] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for locating epileptogenic focus using SEEG electrodes, characterized in that: include: Build an individualized digital twin brain model based on multimodal imaging data and physiological signal data; Based on the digital twin brain model, obtaining possible epileptogenic areas; Based on the possible epileptogenic zone, obtaining an optimal SEEG electrode distribution plan; Based on the optimal SEEG electrode distribution plan, the epileptogenic focus was located.

2. The method for locating epileptogenic focus using SEEG electrodes according to claim 1, characterized in that: The method of constructing an individualized digital twin brain model based on multimodal imaging data and physiological signal data includes: Standardizing and registering the multimodal imaging data, aligning the data of different modalities into the same coordinate system, and obtaining preprocessed data; Segmenting the brain image of the preprocessed data into different anatomical structures, and reconstructing the segmented brain structure data into a three-dimensional brain structure model; Based on the fMRI data in the preprocessed data, identifying the functional connections between different brain regions and obtaining a brain functional network diagram; Based on source reconstruction technology, EEG electrophysiological signals are mapped to a three-dimensional brain structure model to identify the electrical activities of different brain regions; The three-dimensional brain structure model, brain functional network diagram and electrical activities of different brain regions are integrated to obtain a multi-dimensional digital twin brain model.

3. The method for locating epileptogenic focus using SEEG electrodes according to claim 1, characterized in that: The obtaining of possible epileptogenic areas based on the digital twin brain model includes: Perform time-frequency analysis on the EEG signal in each time window to obtain the spectrum characteristics of different time periods and identify the frequency components related to epileptic seizures; Construct a functional connection network between different brain regions, take each brain region as a node, calculate the phase locking value between different brain regions, use the phase locking value between different brain regions as the weight of the edge, obtain the sub-network with abnormally high connection strength (high phase locking value), and determine the potential epileptogenic area; The frequency components related to epileptic seizures and potential epileptogenic areas are input into the machine learning model for training, and the labeled epilepsy data is used for classification to finally predict the possible epileptogenic areas. The final predicted possible epileptogenic zone location is mapped to the digital twin brain model, and based on the anatomical structure and functional connectivity, the marked area in the digital twin brain model is dynamically observed to determine the epileptogenic zone in the digital twin brain model.

4. The method for locating epileptogenic focus using SEEG electrodes according to claim 1, characterized in that: The step of obtaining an optimal SEEG electrode distribution scheme based on the possible epileptogenic zone includes: Based on the epileptogenic zone, determine the target brain area that needs to be monitored; Generate multiple possible distribution paths in the digital twin brain model based on preset constraints and target brain regions; Find the optimal SEEG electrode distribution combination to minimize the number of SEEG electrodes covering the epileptogenic zone.

5. The method for locating epileptogenic focus using SEEG electrodes according to claim 4, characterized in that: The method of finding the optimal SEEG electrode distribution combination to minimize the number of SEEG electrodes covering the epileptogenic zone includes: Based on multiple possible distribution paths, a set of initial electrode distribution combinations is randomly generated as a population. Each distribution combination is regarded as an individual, and the objective function value of each individual is calculated. The higher the objective function, the better the individual. According to the objective function value of the individual, a selection operator is used to select some individuals as parents. Individuals with high objective function values ​​are more likely to be selected. Perform crossover operations on parent individuals, such as single-point crossover or multi-point crossover, to generate offspring individuals; the crossover operation simulates the gene recombination process to generate new distribution combination possibilities; Replace some of the parent individuals with the offspring individuals to form a new population, and calculate the objective function value of each individual in the new population; When the preset maximum number of iterations is reached or the objective function value of the best individual in the population does not change significantly in multiple iterations, the iteration is stopped and the optimal electrode distribution combination is output.

6. The method for locating epileptogenic focus using SEEG electrodes according to claim 5, characterized in that: The objective function min(N) in the optimal electrode distribution combination needs to satisfy C≥C min and S ≥ S min , where N is the number of electrodes, min(N) is the minimum number of electrodes, C is the coverage index of the electrode combination, S is the safety factor, and C min is a pre-set minimum coverage, S min is a pre-set minimum safety factor.

7. The method for locating epileptogenic focus using SEEG electrodes according to claim 1, characterized in that: The method of locating the epileptogenic focus based on the optimal SEEG electrode distribution scheme includes: The space of the digital twin brain model is divided into a number of voxels, the voxels are numbered, and the spatial position of each voxel is recorded; the voxels constitute a potential source space, and each voxel is regarded as a possible source of electrical activity; each voxel is assigned a conductivity, and the conductivity is obtained based on the imaging data of the individual; Based on the geometric shape of the brain, tissue conductivity, and the positional relationship between voxels and SEEG electrodes, the potential generated by each voxel at the electrode position is obtained to obtain a potential distribution matrix, which represents the potential contribution generated by each voxel source at all electrode positions; each row of the positioning distribution matrix corresponds to a SEEG electrode, and each column corresponds to a voxel source in the brain space; Based on the optimal SEEG electrode distribution combination, the EEG signals of the patient during the attack and the inter-attack period are collected, and the EEG signals are pre-processed, including filtering the EEG signals and removing artifacts to obtain the pre-processed EEG signals; Perform feature extraction on the preprocessed EEG signal to obtain the time domain features, frequency domain features and time-frequency features of the EEG signal; A multi-constrained linear equation system is constructed based on time domain features, frequency domain features and time-frequency features, and the regularization strategy is adjusted for the equation system. The equation system constructed based on time domain, frequency domain and time-frequency features and the equation system adjusted by the regularization strategy are jointly solved to obtain the estimated value of the voxel source intensity vector x. The location of the epileptic focus is determined based on the estimated value, and the voxel with the maximum source intensity or the voxel area is determined as the possible location of the epileptic focus.

8. A SEEG electrode localization epileptic focus processing device, used to implement the SEEG electrode localization epileptic focus method of claims 1-7, characterized in that: include: A digital twin brain model construction module, which is used to build an individualized digital twin brain model based on multimodal imaging data and physiological signal data; A possible epileptogenic zone acquisition module, used to acquire the possible epileptogenic zone based on the digital twin brain model; An optimal SEEG electrode distribution scheme acquisition module, used for acquiring an optimal SEEG electrode distribution scheme based on the possible epileptogenic zone; The epileptogenic focus localization module is used to locate the epileptogenic focus based on the optimal SEEG electrode distribution plan.

9. An electronic device, characterized in that: include: one or more processors; Memory; And one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions, which, when executed by the device, enable the device to perform the steps of the SEEG electrode localization epileptogenic focus method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed on a computer, enables the computer to execute the steps of the method for locating epileptogenic focus with SEEG electrodes as described in any one of claims 1 to 7.

Images