Stereotactic electroencephalogram signal analysis method and device, electronic equipment and computer program product

Abstract

The invention discloses a stereotactic electroencephalogram signal analysis method and device, electronic equipment and a computer program product. The method comprises the following steps: acquiring stereotactic electroencephalogram signals of a plurality of channels; performing high-frequency rapid discharge detection on each stereotactic electroencephalogram signal, and determining a channel where the stereotactic electroencephalogram signal with a high-frequency rapid discharge detection result meeting a preset detection condition is located as a key channel; analyzing a causal relationship among the plurality of key channels by using a multivariable autoregression model to obtain a channel adjacency matrix; the key channels are classified according to elements in the channel adjacent matrix, and a channel classification result is obtained and comprises a starting channel, a propagation channel and a termination channel. According to the method, the technical problem that the propagation condition of the brain discharge activity cannot be estimated by the existing stereotactic electroencephalogram signal analysis technology is solved.

Description

Technical Field

[0001] The present invention relates to the field of signal processing, and in particular to a method, device, electronic equipment and computer program product for analyzing stereotactic electroencephalogram (EEG) signals. Background Art

[0002] In the cerebral cortex, the firing activity of neurons is closely related to an individual's behavior and cognitive function. Therefore, studying the firing patterns of neurons is of great significance for understanding brain function. Currently, there are many methods for studying the firing patterns of the cerebral cortex. Among them, stereotaxic electroencephalography (SEEG), as an invasive EEG monitoring method, has become an important tool widely used in clinical and scientific research due to its excellent spatial resolution and directivity.

[0003] SEEG records EEG activity in real time by implanting electrode probes deep within the brain or in specific cortical regions. Due to the flexibility of electrode placement, SEEG can accurately capture discharge patterns in localized brain regions, playing a key role in localizing epileptic foci and studying brain function. However, SEEG signals often contain a significant amount of noise and irrelevant information, necessitating effective signal processing and analysis methods to extract diagnostically valuable features.

[0004] In the existing technology, a variety of methods have been applied to feature extraction and classification analysis of SEEG signals. For example, analysis methods based on information entropy can extract complexity features from EEG signals and achieve preliminary signal recognition and classification. However, these methods mainly focus on static feature extraction of single-channel signals and cannot reveal the dynamic interaction relationship and causal connectivity between multiple channels. They cannot estimate and explain the propagation path and propagation mechanism of brain discharge activity between different channels. Therefore, existing methods have obvious shortcomings in processing multi-source SEEG data and analyzing the neural regulatory mechanisms between different brain regions, and it is difficult to meet the needs of accurate analysis of SEEG signals and clinical decision support.

[0005] In summary, existing technologies have limitations in SEEG signal causal relationship analysis and multi-channel collaborative analysis, and there is still a lack of a systematic and effective solution for analyzing the propagation of brain discharge activity between channels. Summary of the Invention

[0006] The embodiments of the present invention provide a method, device, electronic device and computer program product for analyzing stereotactic EEG signals to at least solve the technical problem that existing stereotactic EEG signal analysis technology cannot estimate the propagation of brain discharge activity.

[0007] According to one aspect of an embodiment of the present invention, a method for analyzing stereotactic EEG signals is provided, comprising: acquiring stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain by a corresponding electrode probe; performing high-frequency rapid discharge detection on each stereotactic EEG signal, and determining that a channel containing the stereotactic EEG signal whose high-frequency rapid discharge detection result meets a preset detection condition is a key channel; and analyzing the causal relationship between the multiple key channels using a multivariate autoregressive model to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct influence of a first channel on a second channel, and the first channel and the second channel are the key channels. The key channels are classified according to the elements in the channel adjacency matrix to obtain a channel classification result, wherein the channel classification result includes: a starting channel, a propagation channel, and an ending channel, the starting channel is the key channel whose out-degree exceeds a first preset condition, the ending channel is the key channel whose in-degree exceeds a second preset condition, the propagation channel is the key channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the key channel is the second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the key channel is the first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0008] Furthermore, a high-frequency rapid discharge detection is performed on each of the stereotactic EEG signals, and the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is determined to be a key channel, including: using a sliding time window of a preset time length to slide on the stereotactic EEG signal according to a preset step length to obtain an energy analysis signal sequence, wherein the energy analysis signal sequence includes: a plurality of energy analysis signals arranged in chronological order, each of the energy analysis signal is the coverage area of ​​the sliding time window on the stereotactic EEG signal after each sliding; performing spectral energy analysis on each of the energy analysis signals to obtain a high- and low-frequency energy ratio, wherein the high- and low-frequency energy ratio is the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal in the same energy analysis signal, the signal frequency of the high-frequency signal is higher than the preset frequency threshold, the signal frequency of the low-frequency signal is not higher than the preset frequency threshold, and the signal energy is the signal Integration of spectral energy density; determining the conversion moment of the stereotactic EEG signal from background slow wave to high-frequency rapid discharge based on the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence, wherein the conversion moment is determined according to the time in the energy analysis signal sequence of the energy analysis signal whose high- and low-frequency energy ratio meets the preset energy ratio condition; determining the channel where the stereotactic EEG signal whose integral ratio meets the preset ratio condition is located as the key channel based on the ratio of the integral of the energy ratio of each stereotactic EEG signal after detecting high-frequency rapid discharge and the integral of the activation delay time, wherein the energy ratio integral is the cumulative value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals after detecting high-frequency rapid discharge, and the activation delay time is the time difference between the conversion moment of each stereotactic EEG signal and the earliest conversion moment among multiple stereotactic EEG signals.

[0009] Furthermore, performing spectral energy analysis on each of the energy analysis signals to obtain the high- and low-frequency energy ratio includes: identifying the high-frequency signal and the low-frequency signal in the energy analysis signal according to the preset frequency threshold, wherein the high-frequency signal includes: a β wave signal and a γ wave signal, and the high-frequency signal includes: an α wave signal and a θ wave signal; detecting the spectral energy density of each candidate signal, wherein the candidate signal is the high-frequency signal or the low-frequency signal; integrating the spectral energy density according to frequency to obtain the signal energy corresponding to each candidate signal; calculating the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal to determine the high- and low-frequency energy ratio.

[0010] Furthermore, based on the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence, determining the conversion moment of the stereotactic electroencephalogram signal from background slow waves to high-frequency rapid discharges includes: determining the average value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence to obtain the average energy ratio; accumulating the deviation between the high- and low-frequency energy ratio corresponding to each energy analysis signal and the average energy ratio according to the order of the energy analysis signals in the energy analysis signal sequence to obtain a cumulative deviation set corresponding to each energy analysis signal, wherein the cumulative deviation set includes: the deviation corresponding to the current energy analysis signal, the deviation corresponding to the historical energy analysis signal, and the order of the historical energy analysis signal in the energy analysis signal sequence is earlier than the current energy analysis signal; when the difference between the deviation corresponding to the energy analysis signal and the smallest deviation in the cumulative deviation set is greater than a preset deviation threshold, determining that the high- and low-frequency energy ratio corresponding to the energy analysis signal meets the preset energy ratio condition, and determining the conversion moment according to the time in the energy analysis signal sequence of the energy analysis signal.

[0011] Furthermore, using a multivariate autoregressive model to analyze the causal relationship between the multiple key channels to obtain a channel adjacency matrix includes: inputting the stereotactic EEG signals of the multiple key channels into the multivariate autoregressive model to obtain a model to be analyzed; performing a discrete-time Fourier transform on the model to be analyzed to obtain a frequency domain coefficient matrix corresponding to each of the key channels, wherein the elements in the frequency domain coefficient matrix represent the amplitude and phase information of the stereotactic EEG signals of the key channels between different frequencies; performing a partial directional coherence analysis on each element in the frequency domain coefficient matrix. A frequency domain matrix is ​​obtained, wherein the elements in the frequency domain matrix represent the frequency domain causal strength of the stereotactic EEG signal of the key channel between different frequencies; the frequency domain causal strength of the stereotactic EEG signals of multiple key channels between target frequency bands is integrated according to frequency to obtain the channel adjacency matrix, wherein the target frequency band is a preset analysis frequency band of the stereotactic EEG signal of the key channel, and each element in the channel adjacency matrix is ​​the integral result of the frequency domain causal strength of the stereotactic EEG signals of the first channel and the second channel.

[0012] Furthermore, the key channels are classified according to the elements in the channel adjacency matrix to obtain the channel classification results, including: calculating the out-degree and the in-degree of each of the stereotactic electroencephalogram signals according to the elements in the channel adjacency matrix; establishing a node to be analyzed corresponding to each key channel in a plane coordinate system, wherein the abscissa of the node to be analyzed is the out-degree, and the ordinate of the node to be analyzed is the in-degree; clustering multiple nodes to be analyzed to obtain multiple clusters; determining the channel classification results based on the multiple clusters, wherein the key channels corresponding to the nodes to be analyzed in the same cluster belong to the same category.

[0013] Furthermore, based on multiple clusters, determining the channel classification result includes: determining the cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is less than the preset in-degree threshold as a first cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster meets the out-degree exceeding the first preset condition; determining the cluster whose out-degree of the cluster center is less than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding the second preset condition; determining the cluster whose out-degree of the cluster center is greater than the preset out-degree threshold The cluster whose out-degree threshold is exceeded and whose in-degree is greater than the preset in-degree threshold is determined as the third cluster, wherein the key channel corresponding to the node to be analyzed in the third cluster meets the out-degree and the in-degree exceeds the key channel of the third preset condition; the signal classification result is determined based on the first cluster, the second cluster and the third cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster is the starting signal, the key channel corresponding to the node to be analyzed in the second cluster is the termination signal, and the key channel corresponding to the node to be analyzed in the third cluster is the propagation signal.

[0014] According to another aspect of an embodiment of the present invention, a device for analyzing stereotactic EEG signals is provided, comprising: an acquisition module for acquiring stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signals of each channel are electrical signals collected from the brain by corresponding electrode probes; a detection module for performing high-frequency rapid discharge detection on each of the stereotactic EEG signals, and determining that the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is a key channel; an analysis module for analyzing the causal relationship between the multiple key channels using a multivariate autoregressive model to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct influence of the first channel on the second channel, and the first channel and the second channel are The key channel; a classification module, used to classify the key channel according to the elements in the channel adjacency matrix to obtain a channel classification result, wherein the channel classification result includes: a starting channel, a propagation channel and an ending channel, the starting channel is the key channel whose out-degree exceeds the first preset condition, the ending channel is the key channel whose in-degree exceeds the second preset condition, the propagation channel is the key channel whose out-degree and in-degree exceed the third preset condition, the out-degree is determined when the key channel is the second channel according to the multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the key channel is the first channel according to the multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0015] According to another aspect of an embodiment of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-mentioned stereotactic electroencephalogram signal analysis method through the computer program.

[0016] According to another aspect of an embodiment of the present invention, a computer program product is provided, comprising computer instructions, which implement the steps of the above-mentioned stereotactic electroencephalogram signal analysis method when executed by a processor.

[0017] The above-mentioned embodiments of the present application, through the high-frequency rapid discharge detection and multivariate autoregressive model analysis of stereotactic EEG signals, can accurately analyze the stereotactic EEG signals based on the causal relationship between multiple stereotactic EEG signals, and accurately divide the stereotactic EEG signals into starting signals, propagation signals and termination signals. This not only improves the accuracy of signal analysis, but also can determine the propagation path information of EEG signals, thereby achieving accurate technical effects on stereotactic EEG signals, and further solving the technical problem that existing stereotactic EEG signal analysis technology cannot estimate the propagation of brain discharge activities. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0019] Figure 1 is a flow chart of a method for analyzing stereotactic electroencephalogram signals according to an embodiment of the present invention;

[0020] Figure 2 is a schematic diagram of a channel adjacency matrix according to an embodiment of the present invention;

[0021] Figure 3 is a schematic diagram of a clustering result of nodes to be analyzed according to an embodiment of the present invention;

[0022] Figure 4 is a schematic diagram of a stereotactic electroencephalogram signal analysis device according to an embodiment of the present invention;

[0023] Figure 5 It is a structural block diagram of a computer terminal according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] According to an embodiment of the present invention, an embodiment of a method for analyzing stereotactic electroencephalogram signals is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] Figure 1 FIG. 1 is a flow chart of a method for analyzing stereotactic EEG signals according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0028] Step S102, obtaining stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain by a corresponding electrode probe;

[0029] Step S104, performing high-frequency rapid discharge detection on each stereotactic EEG signal, and determining the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located as the key channel;

[0030] Step S106: Analyze the causal relationship between multiple key channels using a multivariate autoregressive model to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix represents the direct impact of a first channel on a second channel, and the first channel and the second channel are key channels;

[0031] Step S108: Classify the key channels according to the elements in the channel adjacency matrix to obtain a channel classification result, wherein the channel classification result includes: a starting channel, a propagation channel, and an ending channel. The starting channel is a key channel whose out-degree exceeds a first preset condition, the ending channel is a key channel whose in-degree exceeds a second preset condition, and the propagation channel is a key channel whose out-degree and in-degree exceed a third preset condition. The out-degree is determined, when the key channel is a second channel, based on multiple elements corresponding to the same second channel in the channel adjacency matrix. The in-degree is determined, when the key channel is a first channel, based on multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0032] The above-mentioned embodiments of the present application, through the high-frequency rapid discharge detection and multivariate autoregressive model analysis of stereotactic EEG signals, can determine the propagation of human brain electrical signals based on the causal relationship between multiple stereotactic EEG signals, and accurately divide the channels for collecting stereotactic EEG signals into starting channels, propagation channels and termination channels. Based on the starting channels, propagation channels and termination channels, the starting area of ​​the brain corresponding to the starting channel for generating electrical signals, the propagation area of ​​the transmission signal corresponding to the propagation channel, and the termination area of ​​the termination electrical signal corresponding to the termination channel can be determined, thereby achieving the technical effect of accurately determining the propagation of brain electrical signals based on stereotactic EEG signals, thereby solving the technical problem that the existing stereotactic EEG signal analysis technology cannot estimate the propagation of brain discharge activities.

[0033] The above-mentioned embodiment of the present application uses the energy ratio of the signal and the integral ratio of the activation delay time for preprocessing, which further enhances the reliability of classification. At the same time, by integrating the frequency domain causal strength, a channel adjacency matrix is ​​constructed to achieve quantitative analysis of the causal relationship between signals.

[0034] It should be noted that stereotactic electroencephalography (SEEG) is a technology that records the brain's electrical signals by implanting electrode probes into deep brain structures. It provides more accurate positioning of electrical signals in the brain than traditional electroencephalography (EEG) and magnetoencephalography (MEG), and is of great significance for the diagnosis and treatment of diseases such as epilepsy.

[0035] In the above step S104, high-frequency rapid discharge detection is a key step in identifying epileptic seizures. By detecting abnormal high-frequency signal activities, signals related to epilepsy can be preliminarily screened out, providing a basis for subsequent signal analysis.

[0036] In the above step S106, the multivariate autoregressive model is a statistical model used to analyze the causal relationship between multiple time series. By calculating the time delay and amplitude change between signals, it can reveal the direction and intensity of signal propagation and provide a basis for signal classification.

[0037] In the above step S106, the stereotactic EEG signal of the key channel can be an abnormal electrical signal emitted by the brain in the case of an epileptic seizure. Since the stereotactic EEG signal of the key channel is collected through an electrode probe, the position of the electrode probe in the brain is the position where the brain abnormality is caused by epilepsy.

[0038] In the above step S108, the channel classification result is determined based on the out-degree and in-degree of the stereotactic EEG signal in each channel. The out-degree indicates that the stereotactic EEG signal of a certain channel, as the source, directly affects the stereotactic EEG signals of other channels, and the in-degree indicates that the stereotactic EEG signal of a certain channel, as the receiver, is directly affected by the stereotactic EEG signals of other channels. Through channel classification, the starting, propagation and termination characteristics of the stereotactic EEG signal in each channel can be identified, which is important for understanding the mechanism of epileptic seizures and locating lesions.

[0039] In the above step S108, the stereotactic EEG signal in the starting channel can be the abnormal electrical signal that appears first in the brain during an epileptic seizure. The abnormal electrical signal can propagate in the brain, and the path through the brain during its propagation is the brain path. The channel of the stereotactic EEG signal collected on the brain path is the propagation channel; however, the abnormal electrical signal will not propagate endlessly in the brain, and the position where the brain stops propagating the abnormal electrical signal can be used as the brain termination point, and the electrical signal of the brain termination point can be collected through the termination channel.

[0040] It should be noted that the starting channel, propagation channel and termination channel are all channels for collecting stereotactic EEG signals, and the stereotactic EEG signals in each channel are collected through an electrode probe. Therefore, the position of the electrode probe in the brain is the position where epilepsy causes brain abnormalities, among which the position of the electrode probe corresponding to the starting channel in the brain is the starting position of epilepsy, the position of the electrode probe corresponding to the propagation channel in the brain is the propagation position of epilepsy, and the position of the electrode probe corresponding to the termination channel in the brain is the termination position of epilepsy.

[0041] As an optional embodiment, a high-frequency rapid discharge detection is performed on each stereotactic EEG signal, and determining that the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is a key channel includes: using a sliding time window of a preset time length to slide on the stereotactic EEG signal according to a preset step length to obtain an energy analysis signal sequence, wherein the energy analysis signal sequence includes: multiple energy analysis signals arranged in time sequence, each energy analysis signal is the coverage area of ​​the sliding time window on the stereotactic EEG signal after each sliding; performing spectral energy analysis on each energy analysis signal to obtain a high- and low-frequency energy ratio, wherein the high- and low-frequency energy ratio is the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal in the same energy analysis signal, the signal frequency of the high-frequency signal is higher than the preset frequency threshold, and the signal frequency of the low-frequency signal is not higher than the preset frequency threshold. , the signal energy is the integral of the spectral energy density of the signal; based on the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence, the conversion moment of the stereotactic electroencephalogram signal from the background slow wave to the high-frequency rapid discharge is determined, wherein the conversion moment is determined in the energy analysis signal sequence according to the time of the energy analysis signal whose high- and low-frequency energy ratio meets the preset energy ratio conditions; based on the integral ratio of the energy ratio of each stereotactic electroencephalogram signal after the high-frequency rapid discharge is detected and the integral ratio of the activation delay time, the channel where the stereotactic electroencephalogram signal whose integral ratio meets the preset ratio conditions is located is determined as the key channel, wherein the energy ratio integral is the cumulative value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals after the high-frequency rapid discharge is detected, and the activation delay time is the conversion moment of each stereotactic electroencephalogram signal, and the time difference between it and the earliest conversion moment among the multiple stereotactic electroencephalogram signals.

[0042] It should be noted that the sliding time window technology is a method for dynamically analyzing signals. By moving a window of preset length on the signal at a fixed step size, the characteristics of the signal can be analyzed segment by segment. It is suitable for detecting transient changes in the signal, such as the onset of an epileptic seizure.

[0043] It should be noted that spectral energy analysis is an important step in signal processing. By calculating the energy of the signal at different frequencies, the frequency components of the signal can be identified. High-frequency signals are usually associated with abnormal brain activity, while low-frequency signals represent normal background activity.

[0044] It should be noted that the determination of the conversion moment is achieved by monitoring the change in the energy ratio. When the energy ratio suddenly increases from the background slow wave state and reaches the preset energy ratio condition, it can be considered as the beginning of high-frequency rapid discharge, which provides a time reference for subsequent signal analysis.

[0045] It should be noted that the integral ratio of activation delay time is a quantitative indicator for evaluating the role of signals in epileptic seizures. By comparing the energy ratio integral and activation delay time of the signal, signals that play a key role in epileptic seizures can be screened out, providing a basis for subsequent signal classification.

[0046] The above-described embodiments of the present application, through the use of sliding time windows and spectral energy analysis techniques, can accurately detect high-frequency rapid discharges in stereotactic EEG signals, resolving the difficulty in capturing transient abnormal signals in traditional EEG analysis. By screening stereotactic EEG signals transmitted through key channels using the integral ratio of the conversion moment and activation delay time, the accuracy and specificity of signal analysis are further improved, providing a reliable foundation for subsequent signal classification and lesion localization.

[0047] As an optional embodiment, performing spectral energy analysis on each energy analysis signal to obtain the high-frequency and low-frequency energy ratio includes: identifying the high-frequency signal and the low-frequency signal in the energy analysis signal according to a preset frequency threshold, wherein the high-frequency signal includes: β wave signal and γ wave signal, and the high-frequency signal includes: α wave signal and θ wave signal; detecting the spectral energy density of each candidate signal, wherein the candidate signal is a high-frequency signal or a low-frequency signal; integrating the spectral energy density according to the frequency to obtain the signal energy corresponding to each candidate signal; calculating the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal to determine the high-frequency and low-frequency energy ratio.

[0048] It should be noted that spectral energy density is the energy distribution of a signal at a specific frequency. By detecting energy density, the energy size of the signal at different frequencies can be quantified, providing data support for subsequent energy ratio calculations.

[0049] It should be noted that the calculation of the high-frequency and low-frequency energy ratio is to evaluate the relative intensity of abnormal activity in the signal by comparing the high-frequency signal energy with the low-frequency signal energy. This helps to identify abnormal components in the signal and provide a basis for the diagnosis of epilepsy.

[0050] Optionally, alpha wave signals (8-13 Hz) are associated with a state of relaxation, beta wave signals (13-30 Hz) are associated with a state of alertness, theta wave signals (4-8 Hz) are associated with emotion and memory processing, and gamma wave signals (30-100 Hz) are associated with high concentration and cognitive processing.

[0051] The above-described embodiments of the present application accurately identify beta, gamma, alpha, and theta waves in the signal, and combine this with the integral calculation of the spectral energy density to accurately determine the high- and low-frequency energy ratios of the signal. This solves the problem of inaccurate identification of signal frequency components in traditional EEG analysis. By calculating the energy ratio, the intensity of abnormal activity in the signal can be quantified, providing strong data support for subsequent signal classification and epilepsy diagnosis.

[0052] As an optional embodiment, determining the conversion moment of the stereotactic electroencephalogram signal from background slow wave to high-frequency rapid discharge based on the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence includes: determining the average value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence to obtain the average energy ratio; accumulating the deviation between the high- and low-frequency energy ratio corresponding to each energy analysis signal and the average energy ratio according to the order of the energy analysis signals in the energy analysis signal sequence to obtain a cumulative deviation set corresponding to each energy analysis signal, wherein the cumulative deviation set includes: the deviation corresponding to the current energy analysis signal, the deviation corresponding to the historical energy analysis signal, and the order of the historical energy analysis signal in the energy analysis signal sequence is prior to the current energy analysis signal; when the difference between the deviation corresponding to the energy analysis signal and the smallest deviation in the cumulative deviation set is greater than a preset deviation threshold, determining that the high- and low-frequency energy ratio corresponding to the energy analysis signal meets the preset energy ratio condition, and determining the conversion moment according to the time in the energy analysis signal sequence of the energy analysis signal.

[0053] It should be noted that background slow waves usually refer to low-frequency EEG activity in the brain at rest, while high-frequency rapid discharges represent abnormal high-frequency EEG activity, such as the EEG characteristics during epileptic seizures.

[0054] It should be noted that the average energy ratio is calculated by statistically analyzing the high-frequency and low-frequency energy ratios of all energy analysis signals to obtain a reference value for subsequent deviation calculations.

[0055] It should be noted that the cumulative deviation is calculated by comparing the difference between the high-frequency and low-frequency energy ratios of each energy analysis signal and the average energy ratio, accumulating these differences to form a deviation set for monitoring the changing trend of the energy ratio.

[0056] It should be noted that the conversion moment is determined by monitoring the change in the deviation. When the deviation suddenly increases and exceeds the preset deviation threshold, it can be considered as the beginning of high-frequency rapid discharge, which provides a time reference for subsequent signal analysis.

[0057] The above-described embodiment of the present application accurately identifies the transition moment when stereotactic EEG signals shift from background slow waves to high-frequency rapid discharges by calculating the average energy ratio and cumulative deviation. This solves the problem of traditional EEG analysis in accurately capturing signal state transitions. By monitoring the changing trend of the energy ratio, signal state transitions can be identified in real time, providing a precise temporal reference for subsequent signal classification and lesion localization.

[0058] As an optional embodiment, a multivariate autoregressive model is used to analyze the causal relationship between multiple key channels to obtain a channel adjacency matrix, including: inputting the stereotactic EEG signals of multiple key channels into the multivariate autoregressive model to obtain a model to be analyzed; performing a discrete-time Fourier transform on the model to be analyzed to obtain frequency domain coefficient matrices corresponding to the key channels, wherein the elements in the frequency domain coefficient matrix represent the amplitude and phase information of the stereotactic EEG signals of the key channels between different frequencies; performing partial directional coherence analysis on each element in the frequency domain coefficient matrix to obtain a frequency domain matrix, wherein the elements in the frequency domain matrix represent the frequency domain causal strength of the stereotactic EEG signals of the key channels between different frequencies; integrating the frequency domain causal strength of the stereotactic EEG signals of multiple key channels between the target frequency bands according to the frequency to obtain a channel adjacency matrix, wherein the target frequency band is a pre-set analysis frequency band of the stereotactic EEG signals of the key channels, and each element in the channel adjacency matrix is ​​the integral result of the frequency domain causal strength of the stereotactic EEG signals of the first channel and the second channel.

[0059] It should be noted that the multivariate autoregressive model is a statistical model used to analyze the relationship between multiple time series. By calculating the time delay and amplitude changes between signals, it can reveal the direction and intensity of signal propagation and is suitable for analyzing the dynamic characteristics of complex systems.

[0060] It should be noted that discrete-time Fourier transform is a signal processing technology used to convert time-domain signals into frequency-domain signals. By calculating the amplitude and phase of the signal at different frequencies, the frequency characteristics of the signal can be revealed.

[0061] It should be noted that partial directional coherence analysis is a method to evaluate the causal relationship between signals. By calculating the frequency domain causal strength between signals, it can reveal the direction and intensity of signal propagation and provide a classification basis for the channels that transmit signals.

[0062] It should be noted that the integration of frequency domain causal strength is achieved by integrating the causal strength between signals within the target frequency band to obtain a channel adjacency matrix, which is used to quantify the direct causal relationship between signals transmitted by each channel.

[0063] Optionally, the target frequency band can be: the frequency band of α wave signals, such as 8-13 Hz; or the frequency band of β wave signals, such as 13-30 Hz; or the frequency band of θ wave signals, such as 4-8 Hz; or the frequency band of γ wave signals, such as 30-100 Hz; or a combination of the above frequency bands.

[0064] The above-mentioned embodiments of the present application, by using a multivariate autoregressive model and discrete-time Fourier transform technology, can accurately analyze the causal relationship between stereotactic EEG signals, solving the problem of difficulty in quantifying the causal relationship between signals in traditional EEG analysis. By constructing a channel adjacency matrix, the direct causal relationship between the signals transmitted by each channel can be intuitively demonstrated, providing a quantitative basis for subsequent channel classification and lesion localization.

[0065] As an optional embodiment, the key channels are classified according to the elements in the channel adjacency matrix to obtain the channel classification results, including: calculating the out-degree and in-degree of each stereotactic electroencephalogram signal channel node according to the elements in the channel adjacency matrix; establishing a node to be analyzed corresponding to each key channel in a plane coordinate system, wherein the horizontal coordinate of the node to be analyzed is the out-degree and the vertical coordinate of the node to be analyzed is the in-degree; clustering multiple nodes to be analyzed to obtain multiple clusters; determining the channel classification results based on the multiple clusters, wherein the key channels corresponding to the nodes to be analyzed in the same cluster belong to the same category.

[0066] It should be noted that out-degree and in-degree are key indicators in the channel adjacency matrix. Out-degree indicates the direct influence of the stereotactic EEG signal of a certain channel as the source on the stereotactic EEG signals of other channels, and in-degree indicates that the stereotactic EEG signal of a certain channel as the receiver is directly affected by the stereotactic EEG signals of other channels. By calculating the out-degree and in-degree, the role of the stereotactic EEG signals transmitted by each channel in the signal network can be evaluated.

[0067] It should be noted that the nodes to be analyzed on the plane coordinate system are established by using the out-degree and in-degree of each channel as coordinates to form a signal network to achieve a visual representation of the stereotactic EEG signals in each channel, which is convenient for subsequent cluster analysis.

[0068] It should be noted that cluster analysis is a data mining technology used to group data points with similar characteristics into one category. Through clustering, the stereotactic EEG signals in each channel can be classified according to their role in the signal network, which helps to identify the starting, propagation and termination characteristics of the signal.

[0069] Optionally, the channel classification result is determined based on clustering. Channels in the same cluster have similar out-degree and in-degree, so it can be considered that the stereotactic EEG signals of the channel play a similar role in the signal network, which helps to identify the starting, propagation and termination characteristics of the signal, and provides a basis for subsequent lesion localization.

[0070] The above-described embodiment of the present application accurately classifies key channels by calculating the signal's out-degree and in-degree, combined with cluster analysis on a plane coordinate system. This solves the problem of insufficient classification basis in traditional EEG analysis. The channel classification results can be used to identify the starting, propagation, and termination characteristics of stereotactic EEG signals in each channel, enabling a visual display of signal propagation characteristics.

[0071] As an optional embodiment, based on multiple clusters, determining the channel classification result includes: determining the cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is less than a preset in-degree threshold as a first cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster meets the out-degree exceeding the first preset condition; determining the cluster whose out-degree of the cluster center is less than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding the second preset condition; A cluster whose out-degree of the center is greater than a preset out-degree threshold and whose in-degree is greater than a preset in-degree threshold is determined as the third cluster, wherein the key channel corresponding to the node to be analyzed in the third cluster meets the key channel whose out-degree and in-degree exceed the third preset condition; based on the first cluster, the second cluster and the third cluster, the signal classification result is determined, wherein the key channel corresponding to the node to be analyzed in the first cluster is the starting signal, the key channel corresponding to the node to be analyzed in the second cluster is the termination signal, and the key channel corresponding to the node to be analyzed in the third cluster is the propagation signal.

[0072] It should be noted that the cluster center is the representative point of the cluster. By comparing the out-degree and in-degree of the cluster center, the role of the stereotactic EEG signal in the channel in the signal network can be identified.

[0073] Optionally, stereotactic EEG signals in channels whose out-degree exceeds the first preset condition are generally considered to be starting points of signals because they have a greater impact on other signals in the network while being less affected by them.

[0074] Optionally, stereotactic EEG signals in channels whose in-degree exceeds the second preset condition are considered as termination points of the signals because they are greatly affected by other signals in the network but have little influence on other signals.

[0075] Alternatively, stereotactic EEG signals in channels whose out-degree and in-degree both exceed the third preset condition are considered to be signal propagation points, because they have a greater impact on other signals in the network and are also affected by other signals, which indicates that they play a bridge role in signal propagation.

[0076] The above-mentioned embodiment of the present application accurately identifies the starting, propagation, and termination characteristics of stereotactic EEG signals in a channel by comparing the out-degree and in-degree of the cluster center, thus resolving the difficulty in identifying the propagation path of electrical signals in the brain in traditional EEG analysis. Through the channel classification results, the role of stereotactic EEG signals in each channel in the signal network can be intuitively demonstrated, achieving accurate identification of signal propagation paths and intuitive display of signal action characteristics.

[0077] The present invention also provides an optional embodiment, which provides a method for quantitative analysis of brain networks during epileptic seizures based on stereotactic electroencephalography (SEEG), which implements epileptic seizure network modeling, key node identification, and focal area inference through the following steps, specifically including the following steps:

[0078] Step S21: data preprocessing and segment extraction.

[0079] In step S21, typical data segments covering the entire epileptic seizure process from the pre-ictal stage to the onset of the seizure are extracted from the long-term recorded SEEG signal. This stage combines clinical annotation with a time window sliding algorithm to ensure that the selected segments cover the entire time course of epileptic activity. The specific steps include the following:

[0080] Step S211: filtering and denoising: using a bandpass filter to remove power supply frequency interference and low-frequency drift.

[0081] Step S212: Normalization: Perform Z-score normalization processing on signals of different channels.

[0082] Step S213: Time window selection: A signal window of length T is selected based on clinical seizure annotations.

[0083] Step S214: Data structuring. The data of each channel is constructed into a matrix: Where N is the number of SEEG channels and T is the number of time points.

[0084] Step S22: node selection and epilepsy index calculation.

[0085] In step S22, the extracted SEEG signals are analyzed for an epileptogenicity index (e.g., epileptogenicity index, EI). The high-frequency energy change rate and delay of each SEEG channel at the onset of a seizure are calculated, and channels with an epileptogenicity index higher than a preset threshold are selected as key network nodes, representing potential epileptogenic regions. Specifically, this includes:

[0086] Step S221: high-frequency rapid discharge (Rapid Discharge) detection.

[0087] Step S222: detecting the starting point of rapid discharge.

[0088] Step S223: determining the epileptogenicity index (ie, the integral ratio).

[0089] Step S224, standardization processing.

[0090] Optionally, the goal of high-frequency rapid discharge (Rapid Discharge) detection is to detect the transition moment of the signal from background slow wave to ictal rapid discharge (β / γ band).

[0091] Optionally, the spectral energy density Φ(ω) of each stereotactic EEG signal x(t) is calculated as: Among them, X * (ω) is the Fourier transform of the signal.

[0092] Optionally, the high-low frequency energy ratio (ER) as a sign of rapid discharge is: Among them, E α 、E β 、E γ and E θ Represents the energy of different frequency bands, which can be expressed as E band , where E band =∫ band Φ(ω)dω.

[0093] It should be noted that alpha waves (8-13 Hz) are associated with a state of relaxation, beta waves (13-30 Hz) are associated with a state of alertness, theta waves (4-8 Hz) are associated with emotion and memory processing, and gamma waves (30-100 Hz) are associated with high concentration and cognitive processing.

[0094] Optionally, during actual processing, a sliding window is used to perform spectrum estimation on the discrete signal x[n] to obtain a time-varying ER[n] sequence.

[0095] Optionally, fast discharge onset detection is achieved by applying a cumulative sum change point detection algorithm (Page–Hinkley) to ER[n], specifically including:

[0096] Step S1, calculate the current mean:

[0097] Step S2, construct the cumulative deviation: Where v>0 is the sensitivity adjustment bias term.

[0098] Step S1, define the local minimum u n =min(U1,…,U n ), if: U n -u n>λ, then the rapid discharge detection is triggered, N d is the detection time point (i.e. conversion moment).

[0099] Optionally, the epileptogenicity index (i.e., integral ratio) of the stereotactic EEG signal i of each channel is defined as: the integral of the energy ratio of the channel after detecting a rapid discharge divided by its activation delay time: Among them, N di is the fast discharge detection time of channel i; N0 is the earliest fast discharge time of all channels; ∈ is a constant to avoid division by 0, which is usually set to 1; H is the time window for energy ratio integration (e.g., 5 seconds).

[0100] It should be noted that if a structure does not have rapid discharge, then N di Set as the end time of epileptic seizure, resulting in EI i ≈0.

[0101] Optionally, to facilitate cross-structure or cross-patient comparison, the EI of all channels can be i Normalize: Final The closer the value is to 1, the more likely it is an epileptogenic source.

[0102] Optionally, after determining the epileptogenicity index (i.e., integral ratio) corresponding to the stereotactic EEG signals of multiple channels, the channel where the stereotactic EEG signals in the top 20% of the epileptogenicity index (i.e., integral ratio) are located can be selected as the key channel, or the channel where the stereotactic EEG signals whose epileptogenicity index (i.e., integral ratio) exceeds a preset ratio threshold can be selected as the key channel.

[0103] Step S23: constructing a channel adjacency matrix and causal modeling.

[0104] In step S23, the Directed Transfer Function (DTF) algorithm is used to analyze the direction and strength of information flow between nodes based on the Multivariate Autoregressive (MVAR) model. To reduce the impact of redundancy, a partial correlation metric is introduced to replace the traditional pairwise correlation, improving the specificity and explanatory power of the network structure. Ultimately, a channel adjacency matrix representing the causal relationships between nodes is generated, specifically including the following steps:

[0105] Step S231, multivariate autoregressive model (MVAR).

[0106] set up The SEEG signals of N channels are represented and their dynamic relationships are described using the variable autoregressive model (MVAR): in, is the k-th order coefficient matrix; E(t) is the zero-mean white noise vector; p is the model order.

[0107] Step S232: Dynamic Transfer Function (DTF) derivation.

[0108] Optionally, to obtain the influence relationship between nodes in the frequency domain, the MVAR model needs to be transformed into the frequency domain, specifically including:

[0109] Step S2321, discrete time Fourier transform (DTFT) basis.

[0110] Alternatively, (for a scalar sequence x[n]): Where ω = 2πf / f s ; f is the digital frequency (unit: radians / sample), f s is the sampling rate.

[0111] Optionally, properties of the discrete time Fourier transform (DTFT) include:

[0112] Linear properties:

[0113] Time-shifting properties:

[0114] Alternatively, the DTFT of the vector sequences x(n) and e(n) is defined as:

[0115] It should be noted that the matrix constant A(k) can be directly taken out of the summation symbol.

[0116] Step S2322, perform DTFT on the MVAR equation.

[0117] Alternatively, we can perform DTFT on both sides of the MVAR equation, using the linear and time-shift properties, and we have:

[0118]

[0119] Optionally, the jω ) to the left, and we get:

[0120] Optionally, define the frequency domain coefficient matrix: A(e jω )X(e jω )=E(e jω ),in,

[0121] Step S2323: define the frequency domain transfer function.

[0122] Optionally, define the frequency domain transfer matrix H(e jω ) is: H(e jω )=A -1 (e jω ), X(e jω )=H(e jω )E(e jω ).

[0123] Alternatively, expressed in terms of frequency f, the transfer function can be written as:

[0124] Step S2324, dynamic transfer function (DTF) definition.

[0125] Optionally, the DTF value from node i to node j is defined as:

[0126] It should be noted that DTF considers the “total contribution of the output signal” and is insensitive to collinearity and redundant paths.

[0127] Step S2325: definition of channel adjacency matrix.

[0128] Figure 2 is a schematic diagram of a channel adjacency matrix according to an embodiment of the present invention, such as Figure 2 As shown, the DTF is integrated within the target frequency band (for example, the α wave signal, the frequency band is 8–13 Hz) to obtain the band DTF, and the directional connection matrix is ​​constructed based on it: Where [f1, f2] is the target frequency band; the integral can be numerically expressed as the sum of uniform discrete frequency points (a total of N points): in, is the index set of the corresponding frequency band, and Δf is the frequency point interval.

[0129] Optionally, after obtaining the complete DTF matrix After that, each element Φ ij It represents the directional influence of signal j on signal i within the selected frequency band.

[0130] Optionally, Φ is thinned (e.g., only the top 25% of connections are retained) to construct a directed functional connectivity network. Based on this network, the out-degree, in-degree, and other network metrics of each node can be further calculated for subsequent network structure analysis.

[0131] In the above step S232, the dynamic transfer function (DTF) derivation can also introduce the idea of ​​partial correlation (i.e., partial directional coherence, PDC). To improve the specificity and directional explanatory power of causal inference, DTF is replaced by PDC, which is defined as: in, A ij (f) is the element in the i-th row and j-th column of the frequency domain coefficient matrix.

[0132] It should be noted that PDC measures the "direct impact of variable j on variable i", that is, the causal relationship under the condition of controlling other variables, which essentially introduces partial correlation.

[0133] Optionally, the frequency domain causal strength is integrated within the target frequency band (e.g., 8–13 Hz) to obtain a structured channel adjacency matrix: Among them, [f1,f2] is the target frequency band, and the final channel adjacency matrix is It is a weight matrix of a directional graph, representing the information flow between different SEEG channels (brain regions).

[0134] Step S24: network structure analysis and key node identification.

[0135] In step S24, network analysis and node clustering are performed on the channel adjacency matrix. Combined with network topology indicators (such as out-degree, in-degree, and betweenness centrality), the epileptic network is delineated into seizure onset regions, propagation pathway regions, and seizure termination regions. Key nodes that play a pivotal role in the information propagation pathway are identified to support focal location.

[0136] As an optional embodiment, based on the established directed graph network, network topology analysis is performed to identify the core propagation structure of epileptic seizures, specifically including:

[0137] Step S241: extracting network indicators.

[0138] For each node to be analyzed (i.e., key channel) i, the following indicators are calculated:

[0139] Out-degree: D out i=∑ j A ij;

[0140] In-degree: D in i=∑ j Aji;

[0141] Betweenness Centrality: measures the importance of a node in information dissemination;

[0142] Clustering Coefficient: measures the network density around a node.

[0143] Step S242: network area division.

[0144] Based on topological indicators and temporal characteristics, three epilepsy network sub-areas were divided:

[0145] Onset Zone: The node to be analyzed has a large out-degree, corresponding to the starting signal;

[0146] Propagation Zone: Both out-degree and in-degree are high, and it is in an intermediate position, corresponding to the propagation signal;

[0147] Termination Zone: The in-degree is large, the out-degree is small, and the signal termination is obvious, corresponding to the termination signal.

[0148] Figure 3 FIG. 1 is a schematic diagram of a clustering result of a node to be analyzed according to an embodiment of the present invention. Figure 3 As shown in the figure, the horizontal coordinate of each node to be analyzed is the out-degree, and the vertical coordinate of the node to be analyzed is the in-degree. After clustering, multiple clusters can be obtained. Then, according to the position of each cluster in the two-dimensional coordinate system, the starting channel of the starting area, the propagation channel of the propagation area, and the termination channel of the termination area can be divided.

[0149] The above-mentioned embodiment of the present application is based on quantitative indicators and data-driven models to reduce reliance on subjective interpretation; through causal relationship modeling and key node clustering, it can accurately indicate the starting area of ​​epileptic seizures; this embodiment is applicable to SEEG data analysis of different epilepsy types and electrode layouts, and can serve as an important auxiliary tool for preoperative planning and postoperative efficacy evaluation.

[0150] The above embodiments of the present application can be widely used in neurosurgery, neurology and epilepsy centers for localization of epilepsy foci, formulation of personalized surgical plans and postoperative evaluation, and are particularly suitable for precise treatment path planning for patients with intractable epilepsy.

[0151] According to an embodiment of the present invention, an embodiment of a stereotactic EEG signal analysis device is also provided. It should be noted that the stereotactic EEG signal analysis device can be used to execute the stereotactic EEG signal analysis method in the embodiment of the present invention, and the stereotactic EEG signal analysis method in the embodiment of the present invention can be executed in the stereotactic EEG signal analysis device.

[0152] Figure 4 is a schematic diagram of a stereotactic EEG signal analysis device according to an embodiment of the present invention, such as Figure 4As shown, the device may include: an acquisition module 32 for acquiring stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain by a corresponding electrode probe; a detection module 34 for performing high-frequency rapid discharge detection on each stereotactic EEG signal, and determining that the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is a key channel; an analysis module 36 for using a multivariate autoregressive model to analyze the causal relationship between multiple key channels to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct influence of the first channel on the second channel, and the first channel and the second channel are directly related. The channel is a key channel; the classification module 38 is used to classify the key channels according to the elements in the channel adjacency matrix to obtain a channel classification result, wherein the channel classification result includes: a starting channel, a propagation channel, and an ending channel, the starting channel is a key channel whose out-degree exceeds a first preset condition, the ending channel is a key channel whose in-degree exceeds a second preset condition, the propagation channel is a key channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the key channel is a second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the key channel is a first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0153] It should be noted that the acquisition module 32 in this embodiment can be used to execute step S102 in the embodiment of the present application, the detection module 34 in this embodiment can be used to execute step S104 in the embodiment of the present application, the analysis module 36 in this embodiment can be used to execute step S106 in the embodiment of the present application, and the classification module 38 in this embodiment can be used to execute step S108 in the embodiment of the present application. The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments.

[0154] The above-mentioned embodiments of the present application, through the high-frequency rapid discharge detection and multivariate autoregressive model analysis of stereotactic EEG signals, can accurately analyze the stereotactic EEG signals based on the causal relationship between multiple stereotactic EEG signals, and accurately divide the stereotactic EEG signals into starting signals, propagation signals and termination signals. This not only improves the accuracy of signal analysis, but also can determine the propagation path information of EEG signals, thereby achieving accurate technical effects on stereotactic EEG signals, and further solving the technical problem that existing stereotactic EEG signal analysis technology cannot estimate the propagation of brain discharge activities.

[0155] As an optional embodiment, the detection module includes: a sliding unit, which is used to use a sliding time window of a preset time length to slide on the stereotactic electroencephalogram signal according to a preset step length to obtain an energy analysis signal sequence, wherein the energy analysis signal sequence includes: multiple energy analysis signals arranged in time sequence, each energy analysis signal is the coverage area of ​​the sliding time window on the stereotactic electroencephalogram signal after each sliding; a first analysis unit, which is used to perform spectral energy analysis on each energy analysis signal to obtain a high-frequency and low-frequency energy ratio, wherein the high-frequency and low-frequency energy ratio is the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal in the same energy analysis signal, the signal frequency of the high-frequency signal is higher than the preset frequency threshold, the signal frequency of the low-frequency signal is not higher than the preset frequency threshold, and the signal energy is the integral of the spectral energy density of the signal; a first determination unit, which is used to determine the energy density of the energy analysis signal based on the energy analysis signal. The high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence are used to determine the conversion moment when the stereotactic electroencephalogram signal is converted from background slow wave to high-frequency rapid discharge, wherein the conversion moment is determined according to the time of the energy analysis signal whose high- and low-frequency energy ratio meets the preset energy ratio condition in the energy analysis signal sequence; the second determination unit is used to determine that the channel where the stereotactic electroencephalogram signal whose integral ratio meets the preset ratio condition is the key channel based on the integral ratio of the energy ratio of each stereotactic electroencephalogram signal after the high-frequency rapid discharge is detected and the integral ratio of the activation delay time, wherein the energy ratio integral is the accumulated value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals after the high-frequency rapid discharge is detected, and the activation delay time is the conversion moment of each stereotactic electroencephalogram signal, and the time difference with the earliest conversion moment among the multiple stereotactic electroencephalogram signals.

[0156] As an optional embodiment, the analysis unit includes: an identification subunit, used to identify high-frequency signals and low-frequency signals in the energy analysis signal according to a preset frequency threshold, wherein the high-frequency signals include: β wave signals and γ wave signals, and the high-frequency signals include: α wave signals and θ wave signals; a detection subunit, used to detect the spectral energy density of each candidate signal, wherein the candidate signal is a high-frequency signal or a low-frequency signal; an integration subunit, used to integrate the spectral energy density according to frequency to obtain the signal energy corresponding to each candidate signal; a calculation subunit, used to calculate the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal, and determine the high-low frequency energy ratio.

[0157] As an optional embodiment, the first determination unit includes: a first determination subunit, used to determine the average value of the high-frequency and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence, and obtain the average energy ratio; an accumulation subunit, used to accumulate the deviation between the high-frequency and low-frequency energy ratios corresponding to each energy analysis signal and the average energy ratio according to the order of the energy analysis signals in the energy analysis signal sequence, and obtain a cumulative deviation set corresponding to each energy analysis signal, wherein the cumulative deviation set includes: the deviation corresponding to the current energy analysis signal, the deviation corresponding to the historical energy analysis signal, and the order of the historical energy analysis signal in the energy analysis signal sequence precedes the current energy analysis signal; a second determination subunit, used to determine that the high-frequency and low-frequency energy ratio corresponding to the energy analysis signal meets the preset energy ratio condition when the difference between the deviation corresponding to the energy analysis signal and the smallest deviation in the cumulative deviation set is greater than a preset deviation threshold, and determine the conversion moment according to the time in the energy analysis signal sequence of the energy analysis signal.

[0158] As an optional embodiment, the analysis module includes: an input unit, which is used to input the stereotactic EEG signals of multiple key channels into a multivariate autoregressive model to obtain a model to be analyzed; a Fourier transform unit, which is used to perform discrete-time Fourier transform on the model to be analyzed to obtain frequency domain coefficient matrices corresponding to the key channels, wherein the elements in the frequency domain coefficient matrix represent the amplitude and phase information of the stereotactic EEG signals of the key channels between different frequencies; a second analysis unit, which is used to perform partial directional coherence analysis on each element in the frequency domain coefficient matrix to obtain a frequency domain matrix, wherein the elements in the frequency domain matrix represent the frequency domain causal strength of the stereotactic EEG signals of the key channels between different frequencies; an integration unit, which is used to integrate the frequency domain causal strength of the stereotactic EEG signals of multiple key channels between the target frequency bands according to the frequency to obtain a channel adjacency matrix, wherein the target frequency band is a pre-set analysis frequency band of the stereotactic EEG signals of the key channels, and each element in the channel adjacency matrix is ​​the integral result of the frequency domain causal strength of the stereotactic EEG signals of the first channel and the second channel.

[0159] As an optional embodiment, the classification module includes: a calculation unit, used to calculate the out-degree and in-degree of each stereotactic electroencephalogram signal based on the elements in the channel adjacency matrix; an establishment unit, used to establish the node to be analyzed corresponding to each key channel in the plane coordinate system, wherein the horizontal coordinate of the node to be analyzed is the out-degree, and the vertical coordinate of the node to be analyzed is the in-degree; a clustering unit, used to cluster multiple nodes to be analyzed to obtain multiple clusters; a classification unit, used to determine the channel classification results based on the multiple clusters, wherein the key channels corresponding to the nodes to be analyzed in the same cluster belong to the same category.

[0160] As an optional embodiment, the classification unit includes: a third determining subunit, used to determine the cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is less than a preset in-degree threshold as a first cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster meets the out-degree exceeding the first preset condition; a fourth determining subunit, used to determine the cluster whose out-degree of the cluster center is less than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding the second preset condition; a fifth determining subunit, used to determine the cluster whose out-degree of the cluster center is less than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding the second preset condition; A cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is greater than a preset in-degree threshold is determined as a third cluster, wherein the key channel corresponding to the node to be analyzed in the third cluster meets the key channel whose out-degree and in-degree exceed the third preset condition; a sixth determination subunit is used to determine the signal classification result based on the first cluster, the second cluster and the third cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster is the starting signal, the key channel corresponding to the node to be analyzed in the second cluster is the ending signal, and the key channel corresponding to the node to be analyzed in the third cluster is the propagation signal.

[0161] An embodiment of the present invention may provide an electronic device, which may be a computer terminal, and the computer terminal may be any computer terminal device in a computer terminal group. Optionally, in this embodiment, the computer terminal may also be replaced by a terminal device such as a mobile terminal.

[0162] Optionally, in this embodiment, the computer terminal may be located in at least one network device among a plurality of network devices of a computer network.

[0163] In this embodiment, the computer terminal can execute the program code of the following steps in the method for analyzing stereotactic EEG signals: obtaining stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain by a corresponding electrode probe; performing high-frequency rapid discharge detection on each stereotactic EEG signal, and determining that the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is a key channel; using a multivariate autoregressive model to analyze the causal relationship between multiple key channels to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct influence of the first channel on the second channel, The first channel and the second channel are critical channels; the critical channels are classified according to the elements in the channel adjacency matrix to obtain channel classification results, wherein the channel classification results include: a starting channel, a propagation channel, and an ending channel, the starting channel is a critical channel whose out-degree exceeds a first preset condition, the ending channel is a critical channel whose in-degree exceeds a second preset condition, the propagation channel is a critical channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the critical channel is the second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the critical channel is the first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0164] Figure 5 is a structural block diagram of a computer terminal according to an embodiment of the present invention. Figure 5 As shown, the computer terminal 40 may include: one or more (only one is shown in the figure) processors 42 and a memory 44.

[0165] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the stereotactic EEG signal analysis method and device in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned stereotactic EEG signal analysis method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely located relative to the processor, and these remote memories can be connected to the terminal 40 via a network. Examples of the above-mentioned network include but are not limited to the Internet, corporate intranet, local area network, mobile communication network and combinations thereof.

[0166] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: obtain stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain by a corresponding electrode probe; perform high-frequency rapid discharge detection on each stereotactic EEG signal, and determine that the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is the key channel; use a multivariate autoregressive model to analyze the causal relationship between multiple key channels to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct influence of the first channel on the second channel, and the first channel and the second channel as the key channel; classify the key channels according to the elements in the channel adjacency matrix to obtain channel classification results, wherein the channel classification results include: a starting channel, a propagation channel and an ending channel, the starting channel is a key channel whose out-degree exceeds a first preset condition, the ending channel is a key channel whose in-degree exceeds a second preset condition, the propagation channel is a key channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the key channel is the second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the key channel is the first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0167] Optionally, the processor may also execute the program code of the following steps: using a sliding time window of a preset time length to slide on the stereotactic EEG signal according to a preset step length to obtain an energy analysis signal sequence, wherein the energy analysis signal sequence includes: a plurality of energy analysis signals arranged in chronological order, each energy analysis signal being the coverage area of ​​the sliding time window on the stereotactic EEG signal after each sliding; performing spectral energy analysis on each energy analysis signal to obtain a high- and low-frequency energy ratio, wherein the high- and low-frequency energy ratio is the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal in the same energy analysis signal, the signal frequency of the high-frequency signal is higher than the preset frequency threshold, the signal frequency of the low-frequency signal is not higher than the preset frequency threshold, and the signal energy is the integral of the spectral energy density of the signal; according to the energy analysis signal The high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the signal sequence are used to determine the conversion moment when the stereotactic electroencephalogram signal is converted from background slow wave to high-frequency rapid discharge, wherein the conversion moment is determined according to the time of the energy analysis signal whose high- and low-frequency energy ratio meets the preset energy ratio conditions in the energy analysis signal sequence; according to the ratio of the energy ratio integral of each stereotactic electroencephalogram signal after the high-frequency rapid discharge is detected and the integral ratio of the activation delay time, the channel where the stereotactic electroencephalogram signal whose integral ratio meets the preset ratio conditions is located is determined as the key channel, wherein the energy ratio integral is the cumulative value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals after the high-frequency rapid discharge is detected, and the activation delay time is the conversion moment of each stereotactic electroencephalogram signal, and the time difference between it and the earliest conversion moment among the multiple stereotactic electroencephalogram signals.

[0168] Optionally, the processor may also execute the program code of the following steps: identifying high-frequency signals and low-frequency signals in the energy analysis signal according to a preset frequency threshold, wherein the high-frequency signals include: β wave signals and γ wave signals, and the high-frequency signals include: α wave signals and θ wave signals; detecting the spectral energy density of each candidate signal, wherein the candidate signal is a high-frequency signal or a low-frequency signal; integrating the spectral energy density according to the frequency to obtain the signal energy corresponding to each candidate signal; calculating the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal to determine the high- and low-frequency energy ratio.

[0169] Optionally, the processor may also execute the program code of the following steps: determining the average value of the high-frequency and low-frequency energy ratios corresponding to a plurality of energy analysis signals in the energy analysis signal sequence to obtain the average energy ratio; accumulating the deviation between the high-frequency and low-frequency energy ratios and the average energy ratio corresponding to each energy analysis signal according to the order of the energy analysis signals in the energy analysis signal sequence to obtain a cumulative deviation set corresponding to each energy analysis signal, wherein the cumulative deviation set includes: the deviation corresponding to the current energy analysis signal, the deviation corresponding to the historical energy analysis signal, and the order of the historical energy analysis signal in the energy analysis signal sequence precedes the current energy analysis signal; when the difference between the deviation corresponding to the energy analysis signal and the smallest deviation in the cumulative deviation set is greater than a preset deviation threshold, determining that the high-frequency and low-frequency energy ratio corresponding to the energy analysis signal meets the preset energy ratio condition, and determining the conversion moment according to the time in the energy analysis signal sequence of the energy analysis signal.

[0170] Optionally, the processor may also execute the program code of the following steps: the stereotactic EEG signals of multiple key channels are input into a multivariate autoregressive model to obtain a model to be analyzed; a discrete-time Fourier transform is performed on the model to be analyzed to obtain frequency domain coefficient matrices corresponding to the key channels, wherein the elements in the frequency domain coefficient matrix represent the amplitude and phase information of the stereotactic EEG signals of the key channels between different frequencies; a partial directional coherence analysis is performed on each element in the frequency domain coefficient matrix to obtain a frequency domain matrix, wherein the elements in the frequency domain matrix represent the frequency domain causal strength of the stereotactic EEG signals of the key channels between different frequencies; the frequency domain causal strength of the stereotactic EEG signals of multiple key channels between the target frequency bands is integrated according to the frequency to obtain a channel adjacency matrix, wherein the target frequency band is a preset analysis frequency band of the stereotactic EEG signals of the key channels, and each element in the channel adjacency matrix is ​​the integral result of the frequency domain causal strength of the stereotactic EEG signals of the first channel and the second channel.

[0171] Optionally, the processor may also execute the program code of the following steps: calculating the out-degree and in-degree of each stereotactic EEG signal based on the elements in the channel adjacency matrix; establishing a node to be analyzed corresponding to each key channel in a plane coordinate system, wherein the horizontal coordinate of the node to be analyzed is the out-degree and the vertical coordinate of the node to be analyzed is the in-degree; clustering multiple nodes to be analyzed to obtain multiple clusters; determining the channel classification results based on the multiple clusters, wherein the key channels corresponding to the nodes to be analyzed in the same cluster belong to the same category.

[0172] Optionally, the processor may further execute program code for the following steps: determining a cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is less than a preset in-degree threshold as a first cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster meets the out-degree exceeding a first preset condition; determining a cluster whose out-degree of the cluster center is less than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding a second preset condition; determining a cluster whose out-degree of the cluster center is greater than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a third cluster, wherein the key channel corresponding to the node to be analyzed in the third cluster meets the key channel whose out-degree and in-degree exceed the third preset condition; and determining a signal classification result based on the first cluster, the second cluster, and the third cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster is a starting signal, the key channel corresponding to the node to be analyzed in the second cluster is a termination signal, and the key channel corresponding to the node to be analyzed in the third cluster is a propagation signal.

[0173] It can be understood by those skilled in the art that Figure 5 The structure shown is for illustration only, and the computer terminal may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a handheld computer, a mobile Internet device (MID), a PAD, or other terminal devices. Figure 5 It does not limit the structure of the above electronic device. For example, the computer terminal 40 may also include Figure 5 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 5 Different configurations shown.

[0174] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a computer program. The computer program can be stored in a non-volatile medium. The non-volatile storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0175] The embodiment of the present invention further provides a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium can be used to store the program code executed by the stereotactic EEG signal analysis method provided in the above embodiment.

[0176] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0177] Optionally, in this embodiment, the non-volatile storage medium is configured to store program codes for executing the following steps: obtaining stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal of the brain collected by a corresponding electrode probe; performing high-frequency rapid discharge detection on each stereotactic EEG signal, and determining that the channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets the preset detection conditions is located is a key channel; using a multivariate autoregressive model to analyze the causal relationship between multiple key channels to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct influence of the first channel on the second channel, and the first channel is directly affected by the second channel. The channel and the second channel are critical channels; the critical channels are classified according to the elements in the channel adjacency matrix to obtain channel classification results, wherein the channel classification results include: a starting channel, a propagation channel, and an ending channel, the starting channel is a critical channel whose out-degree exceeds a first preset condition, the ending channel is a critical channel whose in-degree exceeds a second preset condition, the propagation channel is a critical channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the critical channel is the second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the critical channel is the first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

[0178] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: using a sliding time window of a preset time length to slide on the stereotactic electroencephalogram signal according to a preset step size to obtain an energy analysis signal sequence, wherein the energy analysis signal sequence includes: multiple energy analysis signals arranged in time sequence, each energy analysis signal is the coverage area of ​​the sliding time window on the stereotactic electroencephalogram signal after each sliding; performing spectral energy analysis on each energy analysis signal to obtain a high-frequency and low-frequency energy ratio, wherein the high-frequency and low-frequency energy ratio is the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal in the same energy analysis signal, the signal frequency of the high-frequency signal is higher than the preset frequency threshold, and the signal frequency of the low-frequency signal is not higher than the preset frequency threshold, and the signal energy is the integral of the spectral energy density of the signal. ; Based on the high- and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence, the conversion moment of the stereotactic electroencephalogram signal from the background slow wave to the high-frequency rapid discharge is determined, wherein the conversion moment is determined according to the time of the energy analysis signal whose high- and low-frequency energy ratio meets the preset energy ratio conditions in the energy analysis signal sequence; based on the ratio of the energy ratio integral of each stereotactic electroencephalogram signal after the high-frequency rapid discharge is detected and the integral ratio of the activation delay time, the channel where the stereotactic electroencephalogram signal whose integral ratio meets the preset ratio conditions is located is determined as the key channel, wherein the energy ratio integral is the cumulative value of the high- and low-frequency energy ratios corresponding to multiple energy analysis signals after the high-frequency rapid discharge is detected, and the activation delay time is the conversion moment of each stereotactic electroencephalogram signal, and the time difference with the earliest conversion moment among the multiple stereotactic electroencephalogram signals.

[0179] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: identifying high-frequency signals and low-frequency signals in the energy analysis signal according to a preset frequency threshold, wherein the high-frequency signals include: β wave signals and γ wave signals, and the high-frequency signals include: α wave signals and θ wave signals; detecting the spectral energy density of each candidate signal, wherein the candidate signal is a high-frequency signal or a low-frequency signal; integrating the spectral energy density according to frequency to obtain the signal energy corresponding to each candidate signal; calculating the ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal to determine the high- and low-frequency energy ratio.

[0180] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: determining the average value of the high-frequency and low-frequency energy ratios corresponding to multiple energy analysis signals in the energy analysis signal sequence to obtain the average energy ratio; accumulating the deviation between the high-frequency and low-frequency energy ratios corresponding to each energy analysis signal and the average energy ratio according to the order of the energy analysis signals in the energy analysis signal sequence to obtain a cumulative deviation set corresponding to each energy analysis signal, wherein the cumulative deviation set includes: the deviation corresponding to the current energy analysis signal, the deviation corresponding to the historical energy analysis signal, and the order of the historical energy analysis signal in the energy analysis signal sequence precedes the current energy analysis signal; when the difference between the deviation corresponding to the energy analysis signal and the smallest deviation in the cumulative deviation set is greater than a preset deviation threshold, determining that the high-frequency and low-frequency energy ratio corresponding to the energy analysis signal meets the preset energy ratio condition, and determining the conversion moment according to the time in the energy analysis signal sequence of the energy analysis signal.

[0181] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: the stereotactic EEG signals of multiple key channels are input into a multivariate autoregressive model to obtain a model to be analyzed; the model to be analyzed is subjected to a discrete-time Fourier transform to obtain frequency domain coefficient matrices corresponding to the key channels, wherein the elements in the frequency domain coefficient matrix represent the amplitude and phase information of the stereotactic EEG signals of the key channels between different frequencies; a partial directional coherence analysis is performed on each element in the frequency domain coefficient matrix to obtain a frequency domain matrix, wherein the elements in the frequency domain matrix represent the frequency domain causal strength of the stereotactic EEG signals of the key channels between different frequencies; the frequency domain causal strength of the stereotactic EEG signals of multiple key channels between the target frequency bands is integrated according to the frequency to obtain a channel adjacency matrix, wherein the target frequency band is a preset analysis frequency band of the stereotactic EEG signals of the key channels, and each element in the channel adjacency matrix is ​​the integral result of the frequency domain causal strength of the stereotactic EEG signals of the first channel and the second channel.

[0182] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: calculating the out-degree and in-degree of each stereotactic electroencephalogram signal based on the elements in the channel adjacency matrix; establishing a node to be analyzed corresponding to each key channel in a plane coordinate system, wherein the horizontal coordinate of the node to be analyzed is the out-degree and the vertical coordinate of the node to be analyzed is the in-degree; clustering multiple nodes to be analyzed to obtain multiple clusters; determining the channel classification results based on the multiple clusters, wherein the key channels corresponding to the nodes to be analyzed in the same cluster belong to the same category.

[0183] Optionally, in this embodiment, the non-volatile storage medium is configured to store program codes for executing the following steps: determining a cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is less than a preset in-degree threshold as a first cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster meets the out-degree exceeding the first preset condition; determining a cluster whose out-degree of the cluster center is less than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding the second preset condition. ; The cluster whose out-degree of the cluster center is greater than the preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold is determined as the third cluster, wherein the key channel corresponding to the node to be analyzed in the third cluster meets the key channel whose out-degree and in-degree exceed the third preset condition; based on the first cluster, the second cluster and the third cluster, the signal classification result is determined, wherein the key channel corresponding to the node to be analyzed in the first cluster is the starting signal, the key channel corresponding to the node to be analyzed in the second cluster is the termination signal, and the key channel corresponding to the node to be analyzed in the third cluster is the propagation signal.

[0184] An embodiment of the present invention further provides a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, the steps of the stereotactic EEG signal analysis method provided in the above embodiment are implemented.

[0185] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0186] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0187] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0188] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0189] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0190] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned non-volatile storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program code.

[0191] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for analyzing stereotactic electroencephalogram signals, characterized in that: include: Acquiring stereotactic electroencephalogram (EEG) signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain through a corresponding electrode probe; Performing a high-frequency rapid discharge detection on each stereotactic EEG signal, and determining a channel where the stereotactic EEG signal whose high-frequency rapid discharge detection result meets a preset detection condition is located as a key channel; Analyzing the causal relationship between the plurality of key channels using a multivariate autoregressive model to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix is ​​used to represent the direct impact of a first channel on a second channel, and the first channel and the second channel are the key channels; The key channels are classified according to the elements in the channel adjacency matrix to obtain a channel classification result, wherein the channel classification result includes: a starting channel, a propagation channel, and an ending channel, the starting channel is the key channel whose out-degree exceeds a first preset condition, the ending channel is the key channel whose in-degree exceeds a second preset condition, the propagation channel is the key channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the key channel is the second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the key channel is the first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

2. The method according to claim 1, characterized in that Performing high-frequency rapid discharge detection on each stereotactic electroencephalogram signal, and determining as a key channel the channel where the high-frequency rapid discharge detection result of the stereotactic electroencephalogram signal meets the preset detection conditions comprises: Sliding a sliding time window of a preset time length on the stereotactic electroencephalogram signal according to a preset step length to obtain an energy analysis signal sequence, wherein the energy analysis signal sequence includes: a plurality of energy analysis signals arranged in chronological order, each energy analysis signal being a coverage area of ​​the sliding time window on the stereotactic electroencephalogram signal after each sliding; Performing spectral energy analysis on each of the energy analysis signals to obtain a high-frequency and low-frequency energy ratio, wherein the high-frequency and low-frequency energy ratio is a ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal in the same energy analysis signal, the signal frequency of the high-frequency signal is higher than a preset frequency threshold, and the signal frequency of the low-frequency signal is not higher than the preset frequency threshold, and the signal energy is the integral of the spectral energy density of the signal; Determining, based on the high- and low-frequency energy ratios respectively corresponding to a plurality of the energy analysis signals in the energy analysis signal sequence, a transition moment at which the stereotactic electroencephalogram signal transitions from a background slow wave to a high-frequency rapid discharge, wherein the transition moment is determined based on the time in the energy analysis signal sequence when the energy analysis signal whose high- and low-frequency energy ratio meets a preset energy ratio condition; Based on the ratio of the energy ratio integral of each stereotactic EEG signal after detecting high-frequency rapid discharge to the integral ratio of the activation delay time, the channel where the stereotactic EEG signal whose integral ratio meets the preset ratio condition is located is determined to be the key channel, wherein the energy ratio integral is the cumulative value of the high-frequency and low-frequency energy ratios corresponding to multiple energy analysis signals after detecting high-frequency rapid discharge, and the activation delay time is the time difference between the conversion moment of each stereotactic EEG signal and the earliest conversion moment among the multiple stereotactic EEG signals.

3. The method according to claim 2, characterized in that Performing spectrum energy analysis on each of the energy analysis signals to obtain a high-frequency and low-frequency energy ratio includes: According to the preset frequency threshold, the high-frequency signal and the low-frequency signal in the energy analysis signal are identified, wherein the high-frequency signal includes: a beta wave signal and a gamma wave signal, and the high-frequency signal includes: Alpha wave signals and theta wave signals; detecting a spectrum energy density of each candidate signal, wherein the candidate signal is the high-frequency signal or the low-frequency signal; Integrating the spectrum energy density according to frequency to obtain the signal energy corresponding to each candidate signal; The ratio of the signal energy of the high-frequency signal to the signal energy of the low-frequency signal is calculated to determine the high-to-low frequency energy ratio.

4. The method according to claim 2, characterized in that Determining the conversion moment of the stereotactic electroencephalogram signal from the background slow wave to the high-frequency rapid discharge according to the high-frequency and low-frequency energy ratios respectively corresponding to the plurality of energy analysis signals in the energy analysis signal sequence includes: Determining an average value of the high-frequency and low-frequency energy ratios respectively corresponding to a plurality of the energy analysis signals in the energy analysis signal sequence to obtain an average energy ratio; According to the order of the energy analysis signals in the energy analysis signal sequence, the deviations of the high-frequency and low-frequency energy ratios corresponding to each energy analysis signal and the average energy ratio are accumulated to obtain a cumulative deviation set corresponding to each energy analysis signal, wherein the cumulative deviation set includes: the deviation corresponding to the current energy analysis signal and the deviation corresponding to the historical energy analysis signal, the historical energy analysis signal being ordered before the current energy analysis signal in the energy analysis signal sequence; When the difference between the deviation amount corresponding to the energy analysis signal and the smallest deviation amount in the cumulative deviation set is greater than a preset deviation threshold, it is determined that the high-frequency and low-frequency energy ratio corresponding to the energy analysis signal meets the preset energy ratio condition, and the conversion moment is determined according to the time in the energy analysis signal sequence of the energy analysis signal.

5. The method according to claim 1, wherein A multivariate autoregressive model is used to analyze the causal relationship between the multiple key channels, and the channel adjacency matrix obtained includes: Inputting the stereotactic EEG signals of the plurality of key channels into the multivariate autoregressive model to obtain a model to be analyzed; Performing a discrete-time Fourier transform on the model to be analyzed to obtain frequency domain coefficient matrices corresponding to the key channels, wherein elements in the frequency domain coefficient matrices represent amplitude and phase information of the stereotactic electroencephalogram signal of the key channel between different frequencies; Performing partial directional coherence analysis on each element in the frequency domain coefficient matrix to obtain a frequency domain matrix, wherein the elements in the frequency domain matrix represent the frequency domain causal strength of the stereotactic EEG signal of the key channel between different frequencies; The frequency domain causal strength of the stereotactic EEG signals of the multiple key channels between the target frequency bands is integrated according to frequency to obtain the channel adjacency matrix, wherein the target frequency band is a preset analysis frequency band of the stereotactic EEG signals of the key channels, and each element in the channel adjacency matrix is ​​an integral result of the frequency domain causal strength of the stereotactic EEG signals of the first channel and the second channel.

6. The method according to claim 1, characterized in that The key channels are classified according to the elements in the channel adjacency matrix, and the channel classification results obtained include: Calculating the out-degree and the in-degree of each of the stereotactic EEG signals according to the elements in the channel adjacency matrix; On a plane coordinate system, a node to be analyzed corresponding to each key channel is established, wherein the abscissa of the node to be analyzed is the out-degree, and the ordinate of the node to be analyzed is the in-degree; Clustering the plurality of nodes to be analyzed to obtain a plurality of clusters; The channel classification result is determined based on a plurality of clusters, wherein the key channels corresponding to the nodes to be analyzed in the same cluster belong to the same category.

7. The method according to claim 6, characterized in that Determining the channel classification result based on the plurality of clusters includes: Determine the cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is less than the preset in-degree threshold as a first cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster meets the out-degree exceeding the first preset condition; Determine the cluster whose out-degree of the cluster center is less than a preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a second cluster, wherein the key channel corresponding to the node to be analyzed in the second cluster meets the in-degree exceeding a second preset condition; Determining the cluster whose out-degree of the cluster center is greater than a preset out-degree threshold and whose in-degree is greater than the preset in-degree threshold as a third cluster, wherein the key path corresponding to the node to be analyzed in the third cluster meets the key path whose out-degree and in-degree exceed a third preset condition; The signal classification result is determined based on the first cluster, the second cluster, and the third cluster, wherein the key channel corresponding to the node to be analyzed in the first cluster is the start signal, the key channel corresponding to the node to be analyzed in the second cluster is the end signal, and the key channel corresponding to the node to be analyzed in the third cluster is the propagation signal.

8. A stereotactic electroencephalogram signal analysis device, characterized in that: include: an acquisition module, configured to acquire stereotactic EEG signals of multiple channels, wherein the stereotactic EEG signal of each channel is an electrical signal collected from the brain through a corresponding electrode probe; a detection module, configured to perform high-frequency rapid discharge detection on each stereotactic EEG signal, and determine as a key channel the channel where the high-frequency rapid discharge detection result of the stereotactic EEG signal meets a preset detection condition; an analysis module, configured to analyze the causal relationship between the plurality of key channels using a multivariate autoregressive model to obtain a channel adjacency matrix, wherein each element in the channel adjacency matrix represents a direct impact of a first channel on a second channel, and the first channel and the second channel are the key channels; A classification module is used to classify the key channels according to the elements in the channel adjacency matrix to obtain a channel classification result, wherein the channel classification result includes: a starting channel, a propagation channel and an ending channel, the starting channel is the key channel whose out-degree exceeds a first preset condition, the ending channel is the key channel whose in-degree exceeds a second preset condition, the propagation channel is the key channel whose out-degree and in-degree exceed a third preset condition, the out-degree is determined when the key channel is the second channel according to multiple elements corresponding to the same second channel in the channel adjacency matrix, and the in-degree is determined when the key channel is the first channel according to multiple elements corresponding to the same first channel in the channel adjacency matrix.

9. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to execute the stereotactic electroencephalogram signal analysis method according to any one of claims 1 to 7 through the computer program.

10. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by a processor, the steps of the stereotactic electroencephalogram signal analysis method according to any one of claims 1 to 7 are implemented.

Images