Brain tumor real-time identification system and method based on cortical electroencephalogram signals

Abstract

The invention discloses a brain tumor real-time identification system and method based on cortical electroencephalogram signals, and relates to the technical field of medical artificial intelligence, and the method comprises the following steps: obtaining multi-channel ECoG signals in real time; the multi-channel ECoG signals are updated in real time, the multi-channel ECoG signals updated in real time are preprocessed, and the preprocessed multi-channel ECoG signals are obtained; performing channel feature extraction on the ECoG signal of each channel to obtain various ECoG signal features under each channel as channel features; calculating segmentation thresholds of different channel characteristics under each channel based on an automatic threshold segmentation algorithm; carrying out confidence evaluation on the segmentation threshold values of different channel characteristics under each channel; and marking the channel of which the confidence coefficient is smaller than a set tumor marking threshold value as a tumor area. Accurate positioning of a tumor area can be achieved, functional state changes of the cerebral cortex can be reflected in real time, and doctors can conveniently master real-time changes of cortex functions in time.

Description

Technical Field

[0001] This invention belongs to the field of medical artificial intelligence technology, and relates to real-time functional boundary definition in glioma resection surgery, and particularly to a real-time brain tumor identification system and method based on cortical electroencephalogram (EEG) signals. Background Technology

[0002] In neurosurgery, accurate real-time localization of tumor boundaries is crucial for achieving maximum safe resection. Traditional intraoperative navigation techniques primarily rely on preoperative structural images such as MRI, providing surgeons with anatomical references through image registration. However, these techniques suffer from three key drawbacks: First, due to unavoidable brain tissue displacement during surgery, preoperative images can deviate significantly from the actual intraoperative situation; second, structural images cannot reflect real-time changes in the functional state of the cerebral cortex; and most importantly, these techniques cannot provide continuous dynamic monitoring, making it difficult for surgeons to promptly grasp real-time changes in cortical function.

[0003] With the development of minimally invasive neurosurgery, the need for real-time intraoperative functional monitoring is becoming increasingly prominent. This is especially true in glioma surgery in functional areas, where surgeons not only need to accurately identify tumor boundaries but also need to monitor changes in cortical functional status in real time. This real-time monitoring capability has irreplaceable clinical value: first, it allows surgeons to continuously monitor multiple key locations during resection; second, it can dynamically capture changes in cortical electrical activity, providing surgeons with immediate functional assessment feedback, achieving true "real-time navigation." This dynamic monitoring capability is crucial for protecting important functional areas such as language and motor functions.

[0004] Electrocorticometry (ECoG) has become an ideal technique for achieving this goal due to its excellent temporal resolution and functional sensitivity. ECoG signals can reflect changes in the electrical activity of cortical neurons in real time, and its millisecond-level temporal resolution fully meets the real-time monitoring requirements during surgery. Especially in functional area surgery, this real-time monitoring capability is of decisive significance for protecting key cortical functions such as language and motor function.

[0005] However, achieving truly effective intraoperative real-time monitoring requires addressing several key technical challenges. First, the monitoring system must possess sufficient processing speed to achieve millisecond-level signal analysis and feedback. Second, the analysis method needs to remain highly sensitive to characteristic changes in the tumor infiltration area. Finally, the system needs to be adaptable, capable of handling differences in signal characteristics among different patients and brain regions. Meeting these technical requirements directly impacts the final surgical outcome and patient prognosis. Therefore, how to combine cortical electroencephalography (EEG) to achieve rapid detection of tumor boundaries in medical images is a pressing technical problem that needs to be solved.

[0006] Therefore, a real-time brain tumor identification system and method based on cortical electroencephalogram (EEG) signals was developed to solve the above problems. Summary of the Invention

[0007] This invention proposes a real-time brain tumor identification system and method based on cortical electroencephalogram (EEG) signals to solve the problem that significant discrepancies between preoperative imaging and the actual intraoperative situation lead to inaccurate tumor localization, and the inability to reflect real-time changes in the functional state of the cerebral cortex, making it difficult for doctors to grasp real-time changes in cortical function.

[0008] The present invention achieves the above objectives through the following technical solutions:

[0009] This invention provides a real-time brain tumor identification system based on cortical electroencephalogram (EEG) signals, comprising:

[0010] Real-time acquisition of multi-channel ECoG signals;

[0011] The multi-channel ECoG signal is updated in real time, and the updated multi-channel ECoG signal is preprocessed to obtain the preprocessed multi-channel ECoG signal.

[0012] Channel features are extracted from the ECoG signal of each channel to obtain multiple ECoG signal features for each channel as channel features;

[0013] The segmentation threshold for different channel features under each channel is calculated based on an automatic threshold segmentation algorithm;

[0014] Confidence assessment is performed on the segmentation thresholds for different channel features under each channel;

[0015] Channels with confidence levels below a set tumor labeling threshold are labeled as tumor regions.

[0016] Furthermore, the multi-channel ECoG signal is updated in real time, including:

[0017] The acquired multi-channel ECoG signals of a specified duration are stored in a preset ECoG signal data buffer.

[0018] Following a first-in-first-out (FIFO) strategy, the multi-channel ECoG signals in the ECoG signal data buffer are updated in real time.

[0019] Furthermore, the preprocessing of the real-time updated multi-channel ECoG signal includes filtering the ECoG signal in the ECoG signal data buffer.

[0020] Furthermore, the ECoG signal characteristics are at least two of the following: Hjorth mobility parameter, Hjorth complexity parameter, first-order difference, differential entropy, average power, or kurtosis.

[0021] Furthermore, the automatic threshold segmentation algorithm is one or more of the Triangle, Isodata, Li, or Otsu algorithms.

[0022] Furthermore, the formula for calculating the confidence assessment of the segmentation threshold for each channel feature is as follows:

[0023] ;

[0024] In the formula, p represents the tumor confidence level of the channel feature to be evaluated, and F represents the channel feature value to be evaluated. The segmentation threshold is the feature of the channel to be evaluated. This represents the activation function.

[0025] Furthermore, the tumor confidence assessment results under all channel features are integrated to obtain the integrated tumor confidence assessment result. Based on the integrated tumor confidence assessment result, channels with confidence scores lower than the set tumor labeling threshold are marked as tumor regions. The expression for integrating the tumor confidence assessment results under all channel features is as follows:

[0026] ;

[0027] In the formula, This represents the tumor confidence assessment result for the i-th channel feature, where M represents the number of channel feature types to be integrated and evaluated. Integrated assessment results for tumor confidence.

[0028] Furthermore, a real-time brain tumor identification system based on cortical electroencephalogram (EEG) signals also includes real-time generation of tumor images based on channels marked as tumor regions.

[0029] This invention also provides a real-time tumor localization system based on ECoG signals, comprising:

[0030] The acquisition module is used to acquire multi-channel ECoG signals in real time;

[0031] The real-time update and preprocessing module is used to update the multi-channel ECoG signal in real time and preprocess the updated multi-channel ECoG signal to obtain the preprocessed multi-channel ECoG signal.

[0032] The channel feature extraction module is used to extract channel features from the ECoG signal of each channel, and obtain multiple ECoG signal features under each channel as channel features;

[0033] An automatic threshold segmentation module is used to calculate the segmentation threshold for different channel features under each channel based on an automatic threshold segmentation algorithm;

[0034] The confidence assessment module is used to assess the confidence of the segmentation thresholds for different channel features under each channel, and obtain the tumor confidence assessment results for the corresponding channel features.

[0035] The localization module is used to integrate the tumor confidence assessment results under all channel features to obtain the integrated tumor confidence assessment result, and to mark the channels with confidence scores less than a set tumor labeling threshold as tumor regions.

[0036] Furthermore, a real-time tumor localization system based on ECoG signals also includes a tumor imaging unit or a real-time signal acquisition unit. The tumor imaging unit is used to generate tumor images in real time according to channels marked as tumor regions. The real-time signal acquisition unit is used to acquire ECoG signals at several preset locations on the surface of the patient's cerebral cortex, with each preset location corresponding to an ECoG signal channel.

[0037] This invention provides a real-time brain tumor identification system and method based on cortical electroencephalogram (EEG) signals, which can accurately locate the tumor region and reflect the changes in the functional state of the cerebral cortex in real time, making it easier for doctors to grasp the real-time changes in cortical function.

[0038] The beneficial effects of this invention are as follows:

[0039] (1) The present invention acquires multiple features as channel features based on real-time ECoG signals, and then determines the segmentation threshold under the corresponding channel features through an automatic threshold segmentation algorithm, and achieves effective segmentation of the tumor area based on the segmentation threshold; moreover, it can dynamically display the changes in regional electrical activity, which is beneficial for doctors to plan surgery and better protect functional areas.

[0040] (2) The present invention can achieve real-time feedback at the 100ms level, generate 10 frames of dynamic tumor recognition images per second, and continuously capture the dynamic changes of cortical electrical activity caused by surgical operation, which is more than 3,000 times better than MRI in the prior art;

[0041] (3) Compared with the existing MRI technology, the present invention occupies less space (reduced by more than 95%), does not require interruption of the surgical procedure for scanning; and reduces costs by more than 90%, does not require a dedicated shielded operating room, and is easy to train operators. Attached Figure Description

[0042] Figure 1 This is a flowchart of a method for a real-time brain tumor identification system based on cortical electroencephalogram (EEG) signals, as described in an embodiment of this application.

[0043] Figure 2 This is a schematic diagram illustrating the tumor localization effect of a real-time brain tumor identification system based on cortical electroencephalogram (EEG) signals, as described in an embodiment of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0045] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0046] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0047] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0048] like Figure 1 As shown, the present invention discloses a real-time brain tumor identification system based on cortical electroencephalogram (EEG) signals, comprising:

[0049] S1: Real-time acquisition of multi-channel ECoG signals;

[0050] S2: Update the multi-channel ECoG signal in real time, and preprocess the updated multi-channel ECoG signal to obtain the preprocessed multi-channel ECoG signal;

[0051] S3: Extract channel features from the ECoG signal of each channel to obtain multiple ECoG signal features for each channel as channel features;

[0052] S4: Calculate the segmentation threshold for different channel features under each channel based on the automatic threshold segmentation algorithm;

[0053] S5: Confidence assessment of the segmentation thresholds for different channel features under each channel is performed to obtain the tumor confidence assessment results for the corresponding channel features;

[0054] S6: Integrate the tumor confidence assessment results under all channel features to obtain the integrated tumor confidence assessment result, and mark the channels with confidence scores less than the set tumor labeling threshold as tumor regions.

[0055] In one embodiment, real-time updating of the multi-channel ECoG signal includes:

[0056] The acquired multi-channel ECoG signals of a specified time length are stored in a preset ECoG signal data buffer. In this embodiment, a data buffer of 10 seconds is established (the time length is adjustable), that is, 10 seconds of data is used as 1 frame of image.

[0057] Following a first-in-first-out (FIFO) strategy, the multi-channel ECoG signals in the ECoG signal data buffer are updated in real time to maintain the latest 10 seconds of continuous ECoG signals.

[0058] In one embodiment, preprocessing the real-time updated multi-channel ECoG signal includes filtering the ECoG signal in the ECoG signal data buffer.

[0059] In this embodiment, an IIR bandpass filter (4-140Hz) is used for filtering, and an IIR bandstop filter is used to filter out 50Hz power frequency interference.

[0060] In one embodiment, the ECoG signal features are at least two of the following: Hjorth mobility parameter, Hjorth complexity parameter, first-order difference, differential entropy, average power, or kurtosis.

[0061] The Hjorth mobility parameter is expressed as follows:

[0062] ;

[0063] in, The variance function is expressed as:

[0064] ;

[0065] ;

[0066] in, Let N represent the t-th data point of signal y, and N represent the number of data points.

[0067] The first-order difference is represented as:

[0068] ;

[0069] in, , These represent the t-th and t+1-th data points of signal y, respectively. Represents the original ECoG signal at time t The first-order difference;

[0070] The Hjorth complexity parameter is expressed as:

[0071] ;

[0072] in, The first difference of signal y The Hjorth mobility parameter is expressed as:

[0073] ;

[0074] in, The second-order difference is represented as:

[0075] ;

[0076] in, , Let represent the first-order differences of the t-th and t+1-th data points of signal y, respectively. Represents the original ECoG signal at time t The first-order difference;

[0077] For the differential entropy feature, it is expressed as:

[0078] ;

[0079] in, The probability density function of signal y; statistical signals The probability distribution function F(x), where x represents the signal. The value of is obtained by differentiating the probability distribution function F(x). , ;

[0080] The average power characteristic is expressed as:

[0081] ;

[0082] Where N represents the number of data points;

[0083] For kurtosis, it is expressed as:

[0084] ;

[0085] in, and These represent the 4th and 2nd order center distances of signal y, respectively; the k-th order center distance is expressed as:

[0086] .

[0087] In one embodiment, the automatic threshold segmentation algorithm is one or more of the Triangle, Isodata, Li, or Otsu algorithms.

[0088] When the automatic thresholding algorithm is the Triangle algorithm, the segmentation threshold for each channel feature is calculated based on the automatic thresholding algorithm, including:

[0089] Based on the channel feature distribution of all channels, bins are arranged in ascending order of channel feature to construct a normalized histogram;

[0090] Based on the normalized histogram, locate the maximum peak point of the histogram and determine the bin vertex corresponding to the maximum value of the channel feature as the boundary point. Connect the maximum peak point and the boundary point to form a baseline.

[0091] Calculate the vertical distance from the vertex of each bin in the normalized histogram to the baseline, and select the feature value of the channel corresponding to the maximum vertical distance as the segmentation threshold.

[0092] When the automatic thresholding algorithm is the Otsu algorithm, the segmentation threshold for each channel feature is calculated based on the automatic thresholding algorithm, including:

[0093] Based on the channel feature distribution of all channels, bins are arranged in ascending order of channel feature to construct a normalized histogram;

[0094] The inter-class variance is calculated using the following formula, and the segmentation threshold is determined by maximizing the inter-class variance.

[0095] ;

[0096] in, , , and These represent the proportion of non-tumor regions, the proportion of tumor regions, the mean channel feature value of non-tumor regions, and the mean channel feature value of tumor regions, respectively, when th is selected as the segmentation threshold for channel feature values; when When the value is at its maximum, the determined th value is the determined segmentation threshold.

[0097] In one embodiment, when the automatic thresholding algorithm is the Isodata algorithm, the segmentation threshold for channel features under each channel is calculated based on the automatic thresholding algorithm, including:

[0098] Based on the channel feature distribution of all channels, bins are arranged in ascending order of channel feature to construct a normalized histogram;

[0099] Based on the normalized histogram, select the eigenmedian. As an initial threshold, iterate:

[0100] Based on the current threshold Histograms are divided into two categories;

[0101] Calculate the eigenvalues ​​of the two classes and ;

[0102] Update threshold , ;

[0103] like If the value is less than the preset tolerance, the iteration stops.

[0104] The final threshold obtained is the segmentation threshold determined by the algorithm.

[0105] When the automatic thresholding algorithm is the Li algorithm, the segmentation threshold for each channel feature is calculated based on the automatic thresholding algorithm, including:

[0106] Based on the channel feature distribution of all channels, bins are arranged in ascending order of channel feature to construct a normalized histogram;

[0107] For each possible threshold Calculate the probability distribution of the two classes under the current threshold. and ; and based on probability distribution and Calculate the cross-entropy of the two classes. ;

[0108] choose Minimum value This serves as the segmentation threshold determined by the algorithm.

[0109] In one embodiment, the formula for calculating the confidence assessment of the segmentation threshold of channel features under each channel is as follows:

[0110] ;

[0111] In the formula, p represents the tumor confidence level of the channel feature to be evaluated, and F represents the channel feature value to be evaluated. The segmentation threshold is the feature of the channel to be evaluated. This represents the activation function.

[0112] in The function is represented as:

[0113] .

[0114] In one embodiment, the expression for integrating the tumor confidence assessment results under all channel features is as follows:

[0115] ;

[0116] In the formula, This represents the tumor confidence assessment result for the i-th channel feature, where M represents the number of channel feature types to be integrated and evaluated. This represents the integrated assessment results for tumor confidence. In this example, M=3.

[0117] In one embodiment, the method further includes generating tumor images in real time based on channels marked as tumor regions.

[0118] Based on the integrated assessment of tumor confidence, channels with confidence levels below a set tumor labeling threshold (default 50%) are labeled as tumor regions and mapped to electrode channels. Tumor images of the covered areas are then obtained. Figure 2 A schematic diagram illustrating the localization effect of the real-time tumor localization system based on ECoG signals provided in this embodiment is given. When channels 03, 04, 05, 11, and 12 are detected as tumor regions, the tumor identification result is mapped to the electrode channels, and then the tumor image of the covered area is obtained according to the area covered by the electrodes.

[0119] This invention discloses a real-time tumor localization system based on ECoG signals, comprising:

[0120] The acquisition module is used to acquire multi-channel ECoG signals in real time;

[0121] The real-time update and preprocessing module is used to update the multi-channel ECoG signal in real time and preprocess the updated multi-channel ECoG signal to obtain the preprocessed multi-channel ECoG signal.

[0122] The channel feature extraction module is used to extract channel features from the ECoG signal of each channel, and obtain multiple ECoG signal features under each channel as channel features;

[0123] An automatic threshold segmentation module is used to calculate the segmentation threshold for different channel features under each channel based on an automatic threshold segmentation algorithm;

[0124] The confidence assessment module is used to assess the confidence of the segmentation thresholds for different channel features under each channel, and obtain the tumor confidence assessment results for the corresponding channel features.

[0125] The localization module is used to integrate the tumor confidence assessment results under all channel features to obtain the integrated tumor confidence assessment result, and to mark the channels with confidence scores less than a set tumor labeling threshold as tumor regions.

[0126] In one embodiment, when the automatic thresholding algorithm is the Triangle algorithm, the automatic thresholding module includes:

[0127] The feature histogram construction unit is used to construct a normalized histogram by arranging bins according to the channel feature distribution of all channels in ascending order of channel features.

[0128] The baseline construction unit is used to locate the maximum peak point of the histogram based on the normalized histogram, and determine the bin vertex corresponding to the maximum value of the channel feature as the boundary point, and connect the maximum peak point and the boundary point to form the baseline.

[0129] The segmentation threshold calculation unit is used to calculate the vertical distance from the vertex of each bin in the normalized histogram to the baseline, and selects the feature value of the channel corresponding to the maximum vertical distance as the segmentation threshold.

[0130] In one embodiment, when the automatic thresholding algorithm is the Otsu algorithm, the automatic thresholding module includes:

[0131] The feature histogram construction unit is used to construct a normalized histogram by arranging bins according to the channel feature distribution of all channels in ascending order of channel features.

[0132] The segmentation threshold calculation unit is used to calculate the inter-class variance according to the following formula. The segmentation threshold is determined by maximizing the inter-class variance.

[0133] ;

[0134] in, , , and These represent the proportion of non-tumor regions, the proportion of tumor regions, the mean channel feature value of non-tumor regions, and the mean channel feature value of tumor regions, respectively, when th is selected as the segmentation threshold for channel feature values; when When the value is at its maximum, the determined th value is the determined segmentation threshold.

[0135] In one embodiment, a real-time tumor localization system based on ECoG signals further includes a tumor imaging unit, which is used to generate tumor images in real time based on channels marked as tumor regions.

[0136] In one embodiment, a real-time tumor localization system based on ECoG signals further includes a real-time signal acquisition unit, which is used to acquire ECoG signals at several preset locations on the surface of the patient's cerebral cortex, with each preset location corresponding to an ECoG signal channel.

[0137] In one embodiment, a real-time tumor localization system based on ECoG signals further includes a multi-channel ECoG signal real-time update module. The multi-channel ECoG signal real-time update module includes an ECoG signal data buffer and a data update unit. The ECoG signal data buffer is used to store the acquired multi-channel ECoG signals of a specified time length into the preset ECoG signal data buffer. The data update unit is used to update the multi-channel ECoG signals in the ECoG signal data buffer in real time according to a first-in-first-out strategy.

[0138] Specifically, the real-time signal acquisition unit includes several probe electrodes placed directly on the surface of the patient's cerebral cortex at predetermined locations. Each electrode corresponds to one channel. In this embodiment, brain electrical activity is acquired to obtain multi-channel ECoG signals.

[0139] The beneficial effects of this invention are as follows:

[0140] (1) The present invention acquires multiple features as channel features based on real-time ECoG signals, and then determines the segmentation threshold under the corresponding channel features through an automatic threshold segmentation algorithm, and achieves effective segmentation of the tumor area based on the segmentation threshold; moreover, it can dynamically display the changes in regional electrical activity, which is beneficial for doctors to plan surgery and better protect functional areas.

[0141] (2) The present invention can achieve real-time feedback at the 100ms level, generate 10 frames of dynamic tumor recognition images per second, and continuously capture the dynamic changes of cortical electrical activity caused by surgical operation, which is more than 3,000 times better than MRI in the prior art;

[0142] (3) Compared with the existing MRI technology, the present invention occupies less space (reduced by more than 95%), does not require interruption of the surgical procedure for scanning; and reduces costs by more than 90%, does not require a dedicated shielded operating room, and is easy to train operators.

[0143] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals, characterized in that, include: Real-time acquisition of multi-channel ECoG signals; The multi-channel ECoG signal is updated in real time, and the updated multi-channel ECoG signal is preprocessed to obtain the preprocessed multi-channel ECoG signal. Channel features are extracted from the ECoG signal of each channel to obtain multiple ECoG signal features for each channel as channel features; The segmentation threshold for different channel features under each channel is calculated based on an automatic threshold segmentation algorithm; Confidence assessment is performed on the segmentation thresholds for different channel features under each channel; Channels with confidence levels below a set tumor labeling threshold are labeled as tumor regions.

2. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 1, characterized in that, Real-time updates of multi-channel ECoG signals, including: The acquired multi-channel ECoG signals of a specified duration are stored in a preset ECoG signal data buffer. Following a first-in-first-out (FIFO) strategy, the multi-channel ECoG signals in the ECoG signal data buffer are updated in real time.

3. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 1, characterized in that, Preprocessing of the real-time updated multi-channel ECoG signal includes filtering the ECoG signal in the ECoG signal data buffer.

4. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 1, characterized in that, ECoG signal characteristics include at least two of the following: Hjorth mobility parameter, Hjorth complexity parameter, first-order difference, difference entropy, average power, or kurtosis.

5. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 4, characterized in that, The automatic threshold segmentation algorithm is one or more of the Triangle, Isodata, Li, or Otsu algorithms.

6. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 1, characterized in that, The formula for calculating the confidence assessment of the segmentation threshold for each channel feature is as follows: ； In the formula, p represents the tumor confidence level of the channel feature to be evaluated, and F represents the channel feature value to be evaluated. The segmentation threshold is the feature of the channel to be evaluated. This represents the activation function.

7. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 1, characterized in that, The tumor confidence assessment results under all channel features are integrated to obtain the integrated tumor confidence assessment result. Based on the integrated tumor confidence assessment result, channels with confidence scores lower than a set tumor labeling threshold are marked as tumor regions. The expression for integrating the tumor confidence assessment results under all channel features is as follows: ； In the formula, This represents the tumor confidence assessment result for the i-th channel feature, where M represents the number of channel feature types to be integrated and evaluated. Integrated assessment results for tumor confidence.

8. The method for real-time identification of brain tumors based on cortical electroencephalogram (EEG) signals according to claim 7, characterized in that, It also includes generating tumor images in real time based on channels marked as tumor regions.

9. A real-time tumor localization system based on ECoG signals, characterized in that, include: The acquisition module is used to acquire multi-channel ECoG signals in real time; The real-time update and preprocessing module is used to update the multi-channel ECoG signal in real time and preprocess the updated multi-channel ECoG signal to obtain the preprocessed multi-channel ECoG signal. The channel feature extraction module is used to extract channel features from the ECoG signal of each channel, and obtain multiple ECoG signal features under each channel as channel features; An automatic threshold segmentation module is used to calculate the segmentation threshold for different channel features under each channel based on an automatic threshold segmentation algorithm; The confidence assessment module is used to assess the confidence of the segmentation thresholds for different channel features under each channel, and obtain the tumor confidence assessment results for the corresponding channel features. The localization module is used to integrate the tumor confidence assessment results under all channel features to obtain the integrated tumor confidence assessment result, and to mark the channels with confidence scores less than a set tumor labeling threshold as tumor regions.

10. A real-time tumor localization system based on ECoG signals according to claim 9, characterized in that, It also includes a tumor imaging unit or a real-time signal acquisition unit. The tumor imaging unit is used to generate tumor images in real time based on channels marked as tumor regions. The real-time signal acquisition unit is used to acquire ECoG signals at several preset locations on the surface of the patient's cerebral cortex, with each preset location corresponding to an ECoG signal channel.

Images