WO2024259913A1 - 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统 - Google Patents

一种丘脑底核深部脑刺激术电极开机触点选择方法及系统 Download PDF

Info

Publication number
WO2024259913A1
WO2024259913A1 PCT/CN2023/137693 CN2023137693W WO2024259913A1 WO 2024259913 A1 WO2024259913 A1 WO 2024259913A1 CN 2023137693 W CN2023137693 W CN 2023137693W WO 2024259913 A1 WO2024259913 A1 WO 2024259913A1
Authority
WO
WIPO (PCT)
Prior art keywords
mers
amplitude
stn
optimal
module
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/137693
Other languages
English (en)
French (fr)
Inventor
司伟鑫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2024259913A1 publication Critical patent/WO2024259913A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N1/00—Electrotherapy; Circuits therefor
    • A61N1/18—Applying electric currents by contact electrodes
    • A61N1/32—Applying electric currents by contact electrodes alternating or intermittent currents
    • A61N1/36—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation
    • A61N1/3605—Implantable neurostimulators for stimulating central or peripheral nerve system
    • A61N1/3606—Implantable neurostimulators for stimulating central or peripheral nerve system adapted for a particular treatment
    • A61N1/36067—Movement disorders, e.g. tremor or Parkinson disease
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N1/00—Electrotherapy; Circuits therefor
    • A61N1/18—Applying electric currents by contact electrodes
    • A61N1/32—Applying electric currents by contact electrodes alternating or intermittent currents
    • A61N1/36—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation
    • A61N1/3605—Implantable neurostimulators for stimulating central or peripheral nerve system
    • A61N1/36128—Control systems

Definitions

  • the embodiments of the present application relate to the technical field of electroencephalogram (EEG) data processing, and in particular to a method and system for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus.
  • EEG electroencephalogram
  • Parkinson’s Discase is a common progressive neurodegenerative disease that can cause uncontrollable tremors, rigidity, slow walking, lack of coordination and other motor symptoms, which seriously affect the patient’s quality of life.
  • Deep brain stimulation (DBS) therapy has been applied to the field of neuromodulation.
  • DBS therapy electrical stimulation is achieved by an implantable deep brain stimulator.
  • the basic working principle of a deep brain stimulator is to implant stimulating electrodes at the target site of the patient’s brain function (usually the subthalamic nucleus, STN), and connect the electrodes to a pulse generator implanted in the chest through an extension line implanted under the skin.
  • the pulse generator emits stimulating electrical pulses of specified frequency and intensity, which act on the functional areas deep in the patient’s brain to achieve a therapeutic effect.
  • DBS therapy mainly includes MERs signal monitoring, treatment electrode placement, and electrical stimulation of the optimal postoperative start-up contact.
  • the selection of the optimal postoperative contact usually adopts a "trial and error" method, which is judged by observing the patient's muscle tension and movement improvement after electrical stimulation of each contact. This process is somewhat blind, requires a lot of time and effort, and must be completed by experienced programmers.
  • neurosurgeons identify the STN boundary and the optimal position of microelectrode implantation by fusing amplitude sequence and timing features.
  • there is currently no joint analysis algorithm for determining the permanent treatment electrode placement position and selecting the optimal postoperative start-up contact there is currently no joint analysis algorithm for determining the permanent treatment electrode placement position and selecting the optimal postoperative start-up contact.
  • the present application provides a method and system for selecting the electrode start-up contact point for deep brain stimulation of the subthalamic nucleus, so as to solve the problem that the optimal contact selection process after surgery requires repeated experiments and is time-consuming and laborious, thereby improving the postoperative The efficiency of programming and reduce dependence on clinician experience.
  • an embodiment of the present application provides a method for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus, comprising the following steps: first, collecting MERs signals from different brain regions of the patient to be tested; extracting timing features and amplitude features from the MERs signals respectively; fusing the timing features with the amplitude features to obtain an amplitude-frequency perception fusion feature map of the MERs signal; inputting the amplitude-frequency perception fusion feature map into an amplitude-frequency perception fusion network to identify the boundary of the STN and the optimal position of the microelectrode implantation, and determine the boundary of the STN and the placement position of the permanent treatment electrode; analyzing the boundary of the STN and the placement position of the permanent treatment electrode to determine the optimal stimulation contact.
  • time series features and amplitude series features are extracted from MERs signals respectively, including: extracting time series features from MERs signals using a continuous wavelet transform method; and converting MERs signals into GASF visual images using a GASF method, and extracting amplitude series features based on the GASF visual images.
  • the continuous wavelet transform is expressed as:
  • ⁇ is the wavelet basis function
  • a is the scaling coefficient that expands or reduces ⁇ (t) to ⁇ (t/a)
  • a ⁇ 0 is the time shift coefficient of the wavelet position.
  • the GASF matrix is shown in formula (4):
  • the amplitude-frequency sensing fusion network uses a CBAM integrated convolutional neural network model as a classifier to identify the amplitude-frequency sensing fusion feature map to obtain the STN boundary and the placement position of the permanent treatment electrode.
  • the amplitude-frequency-aware deep fusion network uses seven stages to perform full-resolution convolution, each stage includes a convolution layer module and a CBAM module; the last stage includes a fully connected layer and a Softmax function to obtain the output of the amplitude-frequency-aware deep fusion network.
  • the CBAM module includes a channel attention module and a spatial attention module; given an intermediate feature map X as input, the entire CBAM process is expressed as:
  • M C is the channel attention module
  • M S is the spatial attention module
  • X′ represents the features obtained after the channel attention module
  • X′′ is the final output of CBAM.
  • M C is the channel attention module, f AvgPool and f MaxPool are average pooling and maximum pooling respectively. ; MLP represents multi-layer perceptron, ⁇ is the sigmoid function;
  • MS is the spatial attention module and F cov represents the convolution operation.
  • the boundary of the STN and the placement position of the permanent treatment electrode are analyzed to determine the optimal stimulation contact, including: analyzing the boundary of the STN and the placement position of the permanent treatment electrode to obtain the placement strategy of the permanent treatment electrode; determining the optimal stimulation contact based on the placement strategy of the permanent treatment electrode; the placement strategy of the permanent treatment electrode is that the top of the permanent treatment electrode is close to the lower boundary of the STN; the principle of selecting the optimal stimulation contact is to cover the gathering area of the optimal position for the permanent treatment electrode implantation.
  • the embodiment of the present application also provides a system for selecting electrode power-on contact points for deep brain stimulation of the subthalamic nucleus, comprising: a MERs signal acquisition module, a feature extraction module, a feature fusion module, an identification module and an analysis module connected in sequence;
  • the MERs signal acquisition module is used to collect MERs signals from different brain regions of the patient to be tested;
  • the feature extraction module is used to extract timing features and amplitude sequence features from the MERs signal respectively;
  • the feature fusion module is used to fuse the timing features with the amplitude sequence features to obtain an amplitude-frequency perception fusion feature map of the MERs signal;
  • the identification module is used to input the amplitude-frequency perception fusion feature map into the amplitude-frequency perception fusion network, identify the boundary of the STN and the optimal position of the microelectrode implantation, and determine the boundary of the STN and the placement position of the permanent treatment electrode;
  • the analysis module is used to analyze the boundary of the STN and the placement position
  • the present application provides a method and system for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus.
  • the method comprises the following steps: first, collecting MERs signals from different brain regions of a patient to be tested; extracting timing features and amplitude features from the MERs signals respectively; fusing the timing features with the amplitude features to obtain an amplitude-frequency perception fusion feature map of the MERs signals; inputting the amplitude-frequency perception fusion feature map into an amplitude-frequency perception fusion network to identify the boundary of the STN and the optimal position for microelectrode implantation, and determining the boundary of the STN and the permanent position of the STN.
  • the placement of the permanent treatment electrode analyze the boundaries of the STN and the placement of the permanent treatment electrode to determine the optimal stimulation contact.
  • the method for selecting the electrode power-on contact point for deep brain stimulation of the subthalamic nucleus is a method for selecting the optimal contact point of the STN-DBS electrode based on the amplitude-frequency perception deep fusion network.
  • This method uses the amplitude-frequency perception fusion feature and the CBAM integrated CNN model to identify the STN boundary and the optimal position of the microelectrode implantation, guides the placement of the STN-DBS permanent treatment electrode according to the STN boundary, and selects the optimal power-on contact point after surgery according to the optimal position distribution of the microelectrode implantation.
  • the experimental results show that the average accuracy of STN identification in this application is 96.78%, the precision is 97.77%, the sensitivity is 95.26%, and the specificity is 98.07%; the average accuracy of the optimal position identification of the microelectrode implantation is 98.24%, the precision is 98.53%, the sensitivity is 96.05%, and the specificity is 99.31%. All results are better than the current STN and microelectrode implantation optimal position identification algorithm.
  • the optimal contact point selected by the proposed method is consistent with the successful decision-making results of neurosurgeons, which shows that the method proposed in this application has great potential to improve surgical efficiency while reducing dependence on clinician experience.
  • FIG1 is a schematic flow chart of a method for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus provided in one embodiment of the present application;
  • FIG2 is a schematic diagram of a specific flow chart of a method for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus provided in one embodiment of the present application;
  • FIG3 is a schematic diagram of the structure of a CNN model integrated with CBAM provided in an embodiment of the present application.
  • FIG4 is a schematic diagram of the structure of a subthalamic nucleus deep brain stimulation electrode power-on contact point selection system provided by an embodiment of the present application;
  • FIG5 is a schematic diagram of the structure of an electronic device provided in one embodiment of the present application.
  • the selection of the optimal postoperative contact point is usually carried out by a "trial and error” method, which is judged by observing the patient's muscle tension and movement improvement after electrical stimulation of each contact point. This process is somewhat blind, requires a lot of time and energy, and must be completed by experienced programmers.
  • the selection of the optimal contact point after surgery should cover the entire sensorimotor area as much as possible and avoid affecting the substantia nigra.
  • the sensorimotor area of the STN is very small, so it is very difficult for clinicians to locate the boundaries and sensorimotor areas of the STN.
  • the optimal location cluster area for microelectrode implantation identified by MERs signals can usually be considered as the sensorimotor area. According to this principle, doctors can be assisted in accurately identifying the optimal contact point for the treatment electrode.
  • the placement position of the permanent treatment electrode will also affect the selection of the optimal contact point. Therefore, the determination of the placement position of the permanent treatment electrode and the selection of the optimal contact point are indispensable for the success of STN-DBS surgery.
  • a related technology proposes a method for classifying MERs signals using GASF features for automatic identification of the best location for microelectrode implantation.
  • the GASF method converts the one-dimensional time series signal recognition problem into the problem of image classification using artificial intelligence (AI) methods by converting time series data into two-dimensional GASF images.
  • AI artificial intelligence
  • the stripe features displayed by the GASF image are not obvious enough.
  • the time series feature extraction method is not effective. Therefore, the MERs signal recognition algorithm mentioned above only analyzes the time series features or the GASF amplitude series features, which will affect the recognition effect of the MERs signal.
  • the placement of permanent treatment electrodes is determined by neurosurgeons based on MERs.
  • the signal features are identified, and then the optimal start-up contact point is determined through a trial-and-error strategy.
  • the MERs signal category is manually calibrated into five levels by experienced clinicians based on peak, frequency, and amplitude features.
  • the optimal location for microelectrode implantation is marked as the top two levels.
  • the optimal contact point of the electrode is usually determined based on the clustering area of the optimal location for microelectrode implantation.
  • the embodiment of the present application provides a method for selecting the start-up contact points of electrodes for deep brain stimulation of the subthalamic nucleus, comprising the following steps: first, collecting MERs signals from different brain regions of the patient to be tested; extracting the timing features and amplitude features from the MERs signals respectively; fusing the timing features with the amplitude features to obtain the amplitude-frequency perception fusion feature map of the MERs signal; inputting the amplitude-frequency perception fusion feature map into the amplitude-frequency perception fusion network, identifying the boundary of the STN and the optimal position for microelectrode implantation, determining the boundary of the STN and the placement position of the permanent treatment electrode; analyzing the boundary of the STN and the placement position of the permanent treatment electrode to determine the optimal stimulation contact point.
  • the embodiment of the present application provides a method for selecting the start-up contact points of electrodes for deep brain stimulation of the subthalamic nucleus to solve the problem that the process of selecting the best contact point after surgery requires repeated experiments and is time-consuming and labor-intensive, thereby improving the efficiency of postoperative programming and reducing dependence on the experience of clinical physicians.
  • an embodiment of the present application provides a method for selecting electrode power-on contact points for deep brain stimulation of the subthalamic nucleus, comprising the following steps:
  • Step S1 collecting MERs signals from different brain regions of the patient to be tested.
  • Step S2 extracting time series features and amplitude series features from the MERs signal respectively.
  • Step S3 fuse the time series features with the amplitude series features to obtain an amplitude-frequency perception fusion feature map of the MERs signal.
  • Step S4 inputting the amplitude-frequency sensing fusion feature map into the amplitude-frequency sensing fusion network, identifying the boundary of STN and the optimal position for microelectrode implantation, and determining the boundary of STN and the placement position of the permanent treatment electrode.
  • Step S5 Analyze the boundary of STN and the placement position of the permanent treatment electrode to determine the optimal stimulation contact point.
  • this application proposes a method for selecting the best contact point of STN-DBS electrodes based on an amplitude-frequency-aware deep fusion network.
  • a method for selecting the best contact point of STN-DBS electrodes based on an amplitude-frequency-aware deep fusion network First, it is proposed to combine the time series characteristics of the MERs signal with the generated GASF features to obtain the amplitude-frequency-aware fusion features, and then it is proposed that the amplitude-frequency-aware deep fusion network classifies the graded signals recorded by microelectrodes of different frequencies and amplitudes, which can help clinicians effectively identify the STN boundary and the best position for microelectrode implantation, so as to achieve the selection of the best contact point for DBS permanent electrodes.
  • the present application proposes a method for selecting the optimal contact point for STN-DBS electrodes based on an amplitude-frequency-aware deep fusion network.
  • the method first identifies the boundary of the STN to guide the placement of permanent treatment electrodes, and then determines the clustering area of the optimal position for microelectrode implantation, so as to achieve the purpose of accurately selecting the optimal postoperative start-up contact point.
  • the present application uses two-dimensional time series images and GASF images to respectively characterize the high-frequency and high-amplitude features of the MERs time series signals.
  • the method can convert the time series signal classification recognition task into a visual image feature classification task using AI technology for two-dimensional fusion; on the other hand, the present application proposes an amplitude-frequency-aware deep fusion network, which automatically recognizes the graded MERs signals according to different frequencies and amplitudes, and can help clinicians automatically identify the STN boundary and the optimal position for microelectrode implantation. Finally, the present application also proposes a method for automatic decision-making during and after STN-DBS surgery. The method accurately identifies the STN boundary for the determination of the permanent treatment electrode placement position and the selection of the optimal start-up contact point. The experimental results show that neurosurgeons and the method of this article select the same optimal start-up contact point. In addition, the performance of this method in identifying the optimal position of STN and microelectrode implantation is better than the current state-of-the-art methods.
  • the overall flow chart of the method for selecting the optimal contacts of STN-DBS electrodes based on the amplitude-frequency perception deep fusion network proposed in this application is shown in Figure 2.
  • the entire process of the selection method is summarized into two stages: automatic identification of MERs signals based on the amplitude-frequency perception deep fusion network and selection of the optimal stimulation contacts.
  • the difficulties during and after STN-DBS surgery mainly include STN boundary positioning, identification of the optimal implantation position of the microelectrode, and selection of the optimal contacts for permanent treatment electrodes.
  • the amplitude-frequency perception fusion feature map of the MERs signal is obtained by fusing the time series features extracted by CWT with the GASF features.
  • the CBAM integrated CNN model is used to identify the amplitude-frequency perception fusion feature map, which is used to identify the boundary of the STN and the optimal position for microelectrode implantation.
  • the principle of DBS permanent electrode placement is that the electrode head should touch the lower boundary of the STN as much as possible without affecting the substantia nigra, while avoiding adverse reactions to the patient.
  • a comprehensive analysis of the distribution of the position of the permanent treatment electrode and the optimal position of the microelectrode implantation can determine the optimal stimulation contact.
  • Step S1 is the process of collecting MERs signals.
  • MERs signals are collected by an electrophysiological navigation system (NeuroNav 223/12, Alpha Omega Engineering, Israel). Neurosurgeons first use MRI images to preliminarily determine the target location, and plan the microelectrode implantation trajectory in combination with CT scanning technology, and then implant the microelectrode into the STN nucleus according to the preset target location and the planned electrode trajectory.
  • Neurosurgeons first use MRI images to preliminarily determine the target location, and plan the microelectrode implantation trajectory in combination with CT scanning technology, and then implant the microelectrode into the STN nucleus according to the preset target location and the planned electrode trajectory.
  • the MERs measurement method is as follows: (1) Signal acquisition is performed by implanting one or two tungsten microelectrodes; (2) The microelectrode (cannula with a needle core) is inserted approximately 10 mm above the preset target; (3) The microelectrode is further moved in steps of approximately 1 mm until it is inserted approximately 5 mm from the preset target, and a MERs signal is measured at each step; (4) The microelectrode is then advanced in steps of approximately 0.5 mm until it is approximately 4.0 to 5.0 mm below the target, and a MERs signal is measured at each step. The movement of the microelectrode may cause noise interference.
  • MERs data of different lengths may affect the experimental results of signal processing, in order to eliminate noise interference and maintain the stability of the MERs signal, according to the advice of professional neurosurgeons, a MERs signal segment with a duration of 1 second with the least noise was selected at each position.
  • the original MERs signal was preprocessed by professional neurosurgeons using Mapfile Converter software (copyright Alpha Omega 2018, version 5.1.11.0), and then the MERs were cropped and marked using MATLAB software (Mathworks, Inc., Natick, MA) with a non-overlapping sliding window of 1 second. All signals were sampled at 24 kHz and amplified by 1000 times. They were digitally filtered in MATLAB using a 6-pole Butterworth digital bandpass filter from 0.5 kHz to 5 kHz and a notch filter at 60 Hz. All original MERs signals were downsampled by appropriate anti-aliasing filtering. To 12kHz, it can reduce the data sampling points and improve the efficiency of signal recognition.
  • Microelectrodes usually run through the following areas of the brain, mainly including Rt, IC, ZI, H2, STN and SN. Since different brain regions have different electrophysiological signals, the difference in the characteristics of the MERs signals collected from different brain regions can effectively identify the STN boundary and the optimal location for microelectrode implantation. According to the STN boundary position identified by the collected MERs signal and the implantation path detected by the microelectrode, neurosurgeons can determine the implantation trajectory of the permanent treatment electrode, then remove all the microelectrodes, and finally place the four-contact permanent treatment electrode into the STN. In the postoperative stage, clinicians select the optimal contacts based on clinical experience to improve the motor symptoms of Parkinson's patients. This study was approved by the Human Research Ethics Committee of the local hospital. Written informed consent was obtained from all participants after a full explanation of the procedures involved.
  • time series features and amplitude series features are extracted from the MERs signal respectively, including: extracting time series features from the MERs signal using a continuous wavelet transform method; and converting the MERs signal into a GASF visual image using a GASF method, and extracting amplitude series features based on the GASF visual image.
  • the trajectory of microelectrode implantation starts from the prefrontal motor cortex and reaches the substantia nigra.
  • accurate determination of the placement of the treatment electrode is crucial during surgery.
  • the best contact can be selected based on the distribution of the optimal position of the microelectrode implantation during surgery. Therefore, the challenge of STN-DBS lies in identifying the boundary of the STN to determine the placement of the treatment electrode, and identifying the optimal position of the microelectrode implantation to select the optimal contact after surgery.
  • the boundary of the STN and the optimal position of the microelectrode implantation can be identified based on the characteristics of MERs.
  • experienced clinicians usually mark the MERs signals into 5 levels based on the peak, frequency and amplitude characteristics.
  • the MERs located at the optimal position of the microelectrode implantation belong to the highest two levels, and the rest are MERs at non-optimal positions.
  • the identification of the boundary of the STN and the optimal position of the microelectrode implantation can be defined as a typical three-classification problem, that is, the MERs signals come from non-STN, the optimal position in the STN, and the non-optimal position in the STN.
  • MERs signals have complex characteristics such as randomness, multi-components, and non-stationarity
  • converting one-dimensional MERs signals into two-dimensional images can overcome the difficulty of complex one-dimensional feature analysis of MERs.
  • the two most commonly used feature extraction methods for MERs time series image conversion are time series feature extraction and amplitude series feature extraction.
  • continuous wavelet transform has the advantage of analyzing non-stationary signals and can effectively extract time series features.
  • the continuous wavelet transform CWT is expressed as:
  • ⁇ is the wavelet basis function
  • a is the scaling coefficient that expands or reduces ⁇ (t) to ⁇ (t/a)
  • a ⁇ 0 is the time shift coefficient of the wavelet position.
  • Complex Morlet wavelet as a widely used wavelet basis function, has the best time-frequency resolution and the best focusing performance for time series feature extraction in CWT. Therefore, this application selects complex Morlet wavelet (cmor3-3) as the wavelet basis function. Since the longer the scale sequence, the higher the frequency resolution, and a scale sequence that is too long may cause a loss of time resolution, the scale sequence length is set to 256. Since the sampling frequency of the signal is 12kHz, the sampling time is, the total time series length is 1s, and the time-shift coefficient sequence can be obtained along the time axis through the sampling time. The wavelet coefficients are calculated by changing the time-shift coefficients to obtain the final time-frequency feature map. Using the typical time series feature extraction method CWT, the MERs signals located at the optimal position of STN, the non-optimal position of STN, and non-STN can be generated into a 3D time-frequency image.
  • the GASF method is used to convert the MERs signal into a GASF visual image. Given the MERs time series, Y is normalized to the interval [-1,1]. Then the normalized MERs data is converted to polar coordinates, as shown in formula (3):
  • the GASF matrix generates a GASF image with a size of 100 ⁇ 100 ⁇ 3.
  • the GASF images generated by the three types of MERs signals from the best position of STN, the non-best position of STN and non-STN have different texture and stripe characteristics.
  • the amplitude-frequency sensing fusion network uses a CBAM integrated convolutional neural network model as a classifier to identify the amplitude-frequency sensing fusion feature map to obtain the STN boundary and the placement position of the permanent treatment electrode.
  • the time sequence feature map obtained by CWT is fused with the amplitude sequence feature map obtained by the GASF method to obtain the amplitude-frequency perception fusion feature, and the dimension of the obtained amplitude-frequency perception fusion feature map is 100 ⁇ 100 ⁇ 6. Since the attention mechanism can focus on extracting feature information that is conducive to classification in the amplitude-frequency perception fusion feature of the MERs signal and suppress irrelevant neural activity features, this application proposes a CBAM integrated CNN model as a classifier for identifying the amplitude-frequency perception fusion feature map, which can accurately identify the STN boundary and the optimal position of microelectrode implantation.
  • the amplitude-frequency-aware deep fusion network uses seven stages to perform full-resolution convolution, each stage includes a convolutional layer module and a CBAM module; the last stage includes a fully connected layer and a Softmax function to obtain the output of the amplitude-frequency-aware deep fusion network.
  • the amplitude-frequency-aware deep fusion network uses seven stages to perform full-resolution convolution.
  • the amplitude-frequency-aware deep fusion network consists of seven stages, each of which contains a convolutional layer module and a CBAM module.
  • the last stage is a fully connected layer and a Softmax function to obtain the output of the entire network.
  • the Maxpooling layer follows the fourth, fifth, and sixth stages respectively.
  • Each convolutional layer is followed by BN and ReLU.
  • the CBAM module includes a channel attention module and a spatial attention module; given an intermediate feature map X as input, the entire CBAM process is expressed as:
  • M C is the channel attention module
  • M S is the spatial attention module
  • X′ represents the features obtained after the channel attention module
  • X′′ is the final output of CBAM.
  • the input feature map is average pooled and max pooled at the same time, and then passed through a multi-layer perceptron.
  • the multi-layer perceptron contains a shared MLP, and then the output features of the MLP are element-wise summed and then output through a sigmoid activation function.
  • M C is the channel attention module
  • f AvgPool and f MaxPool are average pooling and maximum pooling respectively
  • MLP represents multi-layer perceptron
  • ⁇ is the sigmoid function.
  • the output feature map of the channel attention module is then used as the input feature map of the spatial attention module.
  • the input features are processed by maximum pooling and average pooling, and compressed sampling is performed on the channel dimension.
  • MS is the spatial attention module
  • F cov represents the convolution operation
  • the convolution kernel size is 7 ⁇ 7.
  • the CBAM module can effectively learn emphasized or suppressed features and effectively refine intermediate features.
  • the number of channels of the convolutional layers in stages 1 to 7 are 16, 32, 64, 128, 128, 128, 128, and 128, respectively.
  • the convolution kernel size is 3 ⁇ 3, and the kernel size of the maximum pooling layer is 2 ⁇ 2.
  • the model network is initialized using the "random_uniform" method, and the Adam algorithm is used for network optimization.
  • the learning rate and batch size are set to 1e-2 and 32, respectively.
  • the number of iterations is set to 50.
  • the corresponding samples of the D CWT and D GASF sample sets are spliced to obtain the fusion feature map sample set X, which represents the set of two fusion feature maps generated by the MERs data of 100 people.
  • a block of data in the training set is used as the validation set X val using the ten-fold crossover method, and the remaining data in the training set is updated as a new training set.
  • the obtained new training set and validation set are input into the CBAM enhanced network model for training model parameters.
  • Fifth, in the test process the test set is input into the trained model to obtain the predicted label value y pre , and then the confusion matrix between the predicted result and the actual result is calculated based on the test set label y test , and the classification accuracy, precision and other evaluation results of the algorithm can be obtained to evaluate the model performance.
  • step S5 the boundary of the STN and the implantation position of the permanent treatment electrode are analyzed to determine the optimal stimulation contact, including: analyzing the boundary of the STN and the implantation position of the permanent treatment electrode to obtain the implantation strategy of the permanent treatment electrode; determining the optimal stimulation contact based on the implantation strategy of the permanent treatment electrode; the implantation strategy of the permanent treatment electrode is that the top of the permanent treatment electrode is close to the lower boundary of the STN; the principle of selecting the optimal stimulation contact is to cover the gathering area of the optimal position for implanting the permanent treatment electrode.
  • the selection of the optimal stimulation contacts includes the placement strategy of permanent treatment electrodes and the selection strategy of the optimal postoperative start-up contacts.
  • the placement position of permanent treatment electrodes will seriously affect the selection of the optimal contact point.
  • the reason is that if the permanent treatment electrode is placed too deep, if it exceeds the lower boundary of STN and touches the substantia nigra, it will cause adverse consequences to the patient, and may induce various complications in severe cases. If the electrode is not completely placed in the STN, the coverage of the postoperative start-up programming of the target area is insufficient, and the patient's movement state cannot be completely improved. Before the permanent treatment electrode is placed, it is first necessary to use microelectrodes to detect multiple times to find the boundary of STN and the optimal position of microelectrode implantation.
  • the placement of permanent treatment electrodes is subject to multiple constraints, including the size of the STN, the size of the permanent treatment electrodes, and the area covered by the contact electrical stimulation, as described below:
  • STN size Based on the MERs signal characteristics, the STN boundary range is -1 to 4.5 points, so the actual STN size is 5.5 mm. In addition, the lower boundary of the STN also has obvious differences. Therefore, the placement of permanent treatment electrodes must strictly refer to the actual identified STN area.
  • the size of the permanent treatment electrode is about 7.5 mm long and consists of four 1.5 mm long contacts, spaced 0.5 mm apart. Each contact can be electrically stimulated to program the target area, and the programmable area is also limited.
  • the optimal electrode contact and electrical stimulation should cover the STN as much as possible, especially the sensorimotor area, while avoiding stimulation of the SN to prevent side effects of postoperative programming on the patient.
  • this application summarizes the optimal position for placing the DBS permanent treatment electrode as the top of the electrode being close to the lower border of the STN.
  • the pulse generator is implanted in the chest, and the STN-DBS surgery is completed.
  • the start-up program control that is, to add electrical stimulation to the optimal contact of the treatment electrode to achieve the purpose of improving the patient's motor state.
  • the optimal start-up contact should electrically stimulate the sensorimotor area as much as possible to achieve the best treatment effect.
  • the optimal stimulation contact can be determined according to the distribution of the optimal position of microelectrode implantation, and it should also be noted that electrical stimulation should not affect the SN.
  • the optimal position of microelectrode implantation can be effectively identified by the MERs signal.
  • the principle of selecting the optimal stimulation contact is to cover the clustering area of the optimal position of microelectrode implantation.
  • the clustering area of the optimal position of microelectrode implantation is mainly located in the range of 1 to 2.5 mm of electrode depth, and the electrode contact covered in this range is the contact number 2. Therefore, the selection of the optimal stimulation contact after surgery is based on the clustering area of the optimal position of microelectrode implantation.
  • the method for selecting the electrode power-on contact for deep brain stimulation of the subthalamic nucleus provided by this application was experimentally verified, and the results confirmed that the recognition accuracy of the optimal position of the microelectrode implantation of this application was 97.47%, which can meet the clinical accuracy requirements.
  • clinicians usually select the optimal and suboptimal electrical stimulation contacts as the final candidate contacts.
  • this application can effectively identify the optimal contacts of 20 test patients, and the accuracy can reach 100%, meeting the requirements of precision surgery. Therefore, the selection of this application is consistent with the selection of clinicians, and can provide useful guidance for neurosurgeons.
  • the present application also provides a system for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus, including: a MERs signal acquisition module 101, a feature extraction module 102, a feature fusion module 103, a recognition module 104 and an analysis module 105 connected in sequence;
  • the MERs signal acquisition module 101 is used to collect MERs signals from different brain regions of the patient to be tested;
  • the feature extraction module 102 is used to extract time sequence features and amplitude sequence features from the MERs signal respectively;
  • the feature fusion module 103 is used to fuse the time sequence features with the amplitude sequence features to obtain an amplitude-frequency perception fusion feature map of the MERs signal;
  • the recognition module 104 is used to fuse the amplitude-frequency perception fusion feature map of the MERs signal
  • the known fusion feature map is input into the amplitude-frequency perception fusion network to identify the boundary of STN and the optimal position of microelectrode implantation, and determine the boundary of STN and the
  • another embodiment of the present application provides an electronic device, including: at least one processor 110; and a memory 111 communicatively connected to the at least one processor; wherein the memory 111 stores instructions that can be executed by the at least one processor 110, and the instructions are executed by the at least one processor 110 so that the at least one processor 110 can execute any of the above-mentioned method embodiments.
  • the memory 111 and the processor 110 are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors 110 and the memory 111 together.
  • the bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are all well known in the art, so they are not further described in this article.
  • the bus interface provides an interface between the bus and the transceiver.
  • the transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium.
  • the data processed by the processor 110 is transmitted on a wireless medium through an antenna, and further, the antenna also receives data and transmits the data to the processor 110.
  • the processor 110 is responsible for managing the bus and general processing, and may also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions.
  • the memory 111 may be used to store data used by the processor 110 when performing operations.
  • Another embodiment of the present application relates to a computer-readable storage medium storing a computer program, which implements the above method embodiment when executed by a processor.
  • the embodiment of the present application provides a method and system for selecting electrode power-on contacts for deep brain stimulation of the subthalamic nucleus, which includes the following steps: first, collecting MERs signals from different brain regions of the patient to be tested; extracting timing features and amplitude sequence features from the MERs signals respectively; fusing the timing features with the amplitude sequence features to obtain an amplitude-frequency perception fusion feature map of the MERs signal; inputting the amplitude-frequency perception fusion feature map into an amplitude-frequency perception fusion network to identify the boundary of the STN and the optimal position of the microelectrode implantation, and determining the STN. Analyze the boundaries of STN and the placement of permanent treatment electrodes to determine the optimal stimulation contacts.
  • the method for selecting the electrode power-on contact point for deep brain stimulation of the subthalamic nucleus is a method for selecting the optimal contact point of the STN-DBS electrode based on the amplitude-frequency perception deep fusion network.
  • This method uses the amplitude-frequency perception fusion feature and the CBAM integrated CNN model to identify the STN boundary and the optimal position of the microelectrode implantation, guides the placement of the STN-DBS permanent treatment electrode according to the STN boundary, and selects the optimal power-on contact point after surgery according to the optimal position distribution of the microelectrode implantation.
  • the experimental results show that the average accuracy of STN identification in this application is 96.78%, the precision is 97.77%, the sensitivity is 95.26%, and the specificity is 98.07%; the average accuracy of the optimal position identification of the microelectrode implantation is 98.24%, the precision is 98.53%, the sensitivity is 96.05%, and the specificity is 99.31%. All results are better than the current STN and microelectrode implantation optimal position identification algorithm.
  • the optimal contact point selected by the proposed method is consistent with the successful decision-making results of neurosurgeons, which shows that the method proposed in this application has great potential to improve surgical efficiency while reducing dependence on clinician experience.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

Landscapes

  • Health & Medical Sciences (AREA)
  • Neurology (AREA)
  • Neurosurgery (AREA)
  • Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Hospice & Palliative Care (AREA)
  • Electrotherapy Devices (AREA)

Abstract

一种丘脑底核深部脑刺激术电极开机触点选择方法及系统，涉及脑电数据处理技术领域，方法包括以下步骤：首先，采集待测患者的不同脑区的MERs信号（S1）；从MERs信号中分别提取时序特征与幅序特征（S2）；将时序特征与幅序特征融合，得到MERs信号的幅频感知融合特征图（S3）；将幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置（S4）；对STN的边界以及永久治疗电极的置入位置进行分析，确定最优刺激触点（S5）。解决了术后最佳触点选择过程需要反复试验且耗时费力的难题，提高术后程控的效率并减少对临床医师经验的依赖。

Description

一种丘脑底核深部脑刺激术电极开机触点选择方法及系统 技术领域
本申请实施例涉及脑电数据处理技术领域，特别涉及一种丘脑底核深部脑刺激术电极开机触点选择方法及系统。
背景技术
帕金森(Parkinson’s Discase，PD)是一种常见的进行性神经退行性疾病，可导致不可控的颤抖、强直、行走迟缓、缺乏协调性等运动症状，严重影响患者的生活质量。脑深部电刺激(DBS)疗法已被应用于神经调节领域，在DBS疗法中，电刺激依靠植入式的脑深部刺激器实现。脑深部刺激器的基本工作原理是：在患者大脑功能部位(一般是丘脑底核，STN)靶点植入刺激电极，由埋植于皮下的延长线将电极和埋植在胸前的脉冲发生器相连，脉冲发生器发出指定频率、强度的刺激电脉冲，作用于患者大脑深部的功能区域，达到治疗效果。
DBS疗法主要包括MERs信号监测、治疗电极置入和术后最优开机触点电刺激。目前，术后最优触点的选择通常是采用“试错式”方法，通过观察患者每个触点电刺激之后的肌张力和运动改善情况进行判断，该过程具有一定的盲目性，需要耗费大量的时间和精力，而且还必须由有经验的程控医师来完成。为实现永久治疗电极置入位置的精准定位和术后最优触点的选择，神经外科医生通过将幅序和时序特征融合来识别STN边界和微电极植入的最佳位置。然而，目前还没有对永久治疗电极置入位置的确定和术后最优开机触点的选择的联合分析算法。
发明内容
本申请实施例提供一种丘脑底核深部脑刺激术电极开机触点选择方法及系统，以解决术后最佳触点选择过程需要反复试验且耗时费力的难题，提高术后 程控的效率并减少对临床医师经验的依赖。
为解决上述技术问题，第一方面，本申请实施例提供一种丘脑底核深部脑刺激术电极开机触点选择方法，包括以下步骤：首先，采集待测患者的不同脑区的MERs信号；从MERs信号中分别提取时序特征与幅序特征；将时序特征与幅序特征融合，得到MERs信号的幅频感知融合特征图；将幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置；对STN的边界以及永久治疗电极的置入位置进行分析，确定最优刺激触点。
在一些示例性实施例中，从MERs信号中分别提取时序特征与幅序特征，包括：采用连续小波变换方法，从MERs信号中提取时序特征；以及采用GASF方法将MERs信号转换为GASF视觉图像，并基于GASF视觉图像，提取幅序特征。
在一些示例性实施例中，连续小波变换表示为：

公式(1)和公式(2)中，ψ为小波基函数，a为尺度系数将ψ(t)扩展或缩小到ψ(t/a)，a≠0，b为小波位置的时移系数。
在一些示例性实施例中，采用GASF方法将所述MERs信号转换为GASF视觉图像，包括：将MERs时间序列归一化至区间[-1,1]，得到归一化后的MERs数据；MERs时间序列表示为Y＝{y1，y2，...，yn}；将归一化后的MERs数据转换为极坐标；基于极坐标，得到GASF矩阵；基于GASF矩阵，生成GASF视觉图像；极坐标如公式(3)所示：
所述GASF矩阵如公式(4)所示：
其中，角度φ和半径r利用归一化的MERs的值和相应的时间戳t来表示，其中N是正则化跨度的常数因子。
在一些示例性实施例中，幅频感知融合网络采用CBAM集成的卷积神经网络模型为分类器，对幅频感知融合特征图进行识别，得到STN边界以及永久治疗电极的置入位置。
在一些示例性实施例中，幅频感知深度融合网络采用七个阶段执行全分辨率的卷积，每个阶段包括一个卷积层模块和一个CBAM模块；最后一个阶段包括一个全连接层和一个Softmax函数，用来获得幅频感知深度融合网络的输出。
在一些示例性实施例中，CBAM模块包括通道注意力模块和空间注意力模块；给定一个中间特征映射X作为输入，整个CBAM过程表示为：
其中，MC为通道注意力模块，MS为空间注意力模块，表示逐个元素相乘，X′表示通过通道注意力模块后获得的特征，X″为CBAM的最终输出。
在一些示例性实施例中，所述通道注意力模块的计算公式如下所示：
MC(X)＝σ(MLP(fAvgPool(X))+MLP(fMaxPool(X)))     (6)
其中，MC为通道注意力模块，fAvgPool和fMaxPool分别为平均池化和最大池 化；MLP表示多层感知机，σ为sigmoid函数；
空间注意力模块的计算公式如下所示：
MS(X′)＝σ(Fcov([fAvgPool(X)；fMaxPool(X)]))      (7)
其中，MS为空间注意力模块，Fcov表示卷积运算。
在一些示例性实施例中，对STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点，包括：对STN的边界以及永久治疗电极的置入位置进行分析，得到永久治疗电极的置入策略；基于永久治疗电极的置入策略，确定最优刺激触点；永久治疗电极的置入策略为永久治疗电极的顶端靠近STN的下边界；最优刺激触点的选择的原则是覆盖永久治疗电极植入的最佳位置的聚集区。
第二方面，本申请实施例还提供了一种丘脑底核深部脑刺激术电极开机触点选择系统，包括：依次连接的MERs信号采集模块、特征提取模块、特征融合模块、识别模块以及分析模块；MERs信号采集模块用于采集待测患者的不同脑区的MERs信号；特征提取模块用于从所述MERs信号中分别提取时序特征与幅序特征；特征融合模块用于将所述时序特征与所述幅序特征融合，得到MERs信号的幅频感知融合特征图；识别模块用于将所述幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置；分析模块用于对所述STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点。
本申请实施例提供的技术方案至少具有以下优点：
本申请实施例提供一种丘脑底核深部脑刺激术电极开机触点选择方法及系统，该方法包括以下步骤：首先，采集待测患者的不同脑区的MERs信号；从MERs信号中分别提取时序特征与幅序特征；将时序特征与幅序特征融合，得到MERs信号的幅频感知融合特征图；将幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永 久治疗电极的置入位置；对STN的边界以及永久治疗电极的置入位置进行分析，确定最优刺激触点。
本申请提供的丘脑底核深部脑刺激术电极开机触点选择方法，是基于幅频感知深度融合网络的STN-DBS电极最优触点的选择方法。该方法将幅频感知融合特征和CBAM集成CNN模型用于STN边界和微电极植入的最佳位置的识别，根据STN边界指导STN-DBS永久治疗电极的置入，依据微电极植入的最佳位置分布选择术后最佳的开机触点。实验结果表明，本申请STN识别的平均准确率为96.78％，精确率为97.77％，灵敏度为95.26％，特异度为98.07％；微电极植入的最佳位置识别的平均准确率为98.24％，精确率为98.53％，灵敏度为96.05％，特异度为99.31％，所有结果均优于目前的STN和微电极植入的最佳位置识别算法。所提出的方法选择的最优触点与神经外科医生的成功决策结果一致，这表明本申请所提出的方法在提高手术效率的同时减少对临床医生经验的依赖有很大的潜力。
附图说明
一个或多个实施例通过与之对应的附图中的图片进行示例性说明，这些示例性说明并不构成对实施例的限定，除非有特别申明，附图中的图不构成比例限制。
图1为本申请一实施例提供的一种丘脑底核深部脑刺激术电极开机触点选择方法的流程示意图；
图2为本申请一实施例提供的一种丘脑底核深部脑刺激术电极开机触点选择方法的具体流程示意图；
图3为本申请一实施例提供的CBAM集成的CNN模型结构示意图；
图4为本申请一实施例提供的一种丘脑底核深部脑刺激术电极开机触点选择系统的结构示意图；
图5为本申请一实施例提供的一种电子设备的结构示意图。
具体实施方式
由背景技术可知，目前术后最优触点的选择通常是采用“试错式”方法，通过观察患者每个触点电刺激之后的肌张力和运动改善情况进行判断，该过程具有一定的盲目性，需要耗费大量的时间和精力，而且还必须由有经验的程控医师来完成。
术后最优触点的选择应该尽可能的覆盖整个感觉运动区且避免影响到黑质。然而STN的感觉运动区非常小，因此定位STN的边界和感觉运动区对临床医生来说具有很大的难度。MERs信号识别的微电极植入的最佳位置聚集区通常可以被认为是感觉运动区，根据这个原则可以辅助医生准确的识别治疗电极的最优触点。此外，永久治疗电极的置入位置也会影响最优触点的选择。因此，永久治疗电极的置入位置的确定和最优触点的选择对于STN-DBS手术的成功缺一不可。
临床医生选择最优触点通常还可以借助于MERs信号识别的微电极植入的最佳位置。一相关技术提出了一种利用GASF特征对MERs信号进行分类，用于微电极植入的最佳位置的自动识别。GASF方法是通过将时间序列数据转换为二维的GASF图像，可以将一维时间序列信号识别问题转换为通过人工智能(Artificial intelligence，AI)的方法对图像进行分类的问题。对于部分高频高幅特征的MERs信号，GASF图像显示的条纹特征不够明显，然而对峰值不明显的MERs信号，时间序列特征提取方法效果不佳。因此，上面提到的MERs信号识别算法只分析时序特征或者GASF幅序特征都会影响MERs信号的识别效果。
为实现永久治疗电极置入位置的精准定位和术后最优触点的选择，神经外科医生通过将幅序和时序特征融合来识别STN边界和微电极植入的最佳位置。然而，目前还没有对永久电极置入位置的确定和术后最优开机触点的选择的联合分析算法。临床上，永久治疗电极的置入位置是由神经外科医生依据MERs 信号特征进行识别，然后通过试错策略确定最优开机触点。MERs信号类别是由经验丰富的临床医生根据峰值、频率和振幅特征等手动标定为五个等级，微电极植入的最佳位置被标记为最高两个等级，电极的最优触点通常根据微电极植入的最佳位置的聚集区确定。
为解决上述技术问题，本申请实施例提供一种丘脑底核深部脑刺激术电极开机触点选择方法，包括以下步骤：首先，采集待测患者的不同脑区的MERs信号；从MERs信号中分别提取时序特征与幅序特征；将时序特征与幅序特征融合，得到MERs信号的幅频感知融合特征图；将幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置；对STN的边界以及永久治疗电极的置入位置进行分析，确定最优刺激触点。本申请实施例通过提供一种丘脑底核深部脑刺激术电极开机触点选择方法，以解决术后最佳触点选择过程需要反复试验且耗时费力的难题，提高术后程控的效率并减少对临床医师经验的依赖。
下面将结合附图对本申请的各实施例进行详细的阐述。然而，本领域的普通技术人员可以理解，在本申请各实施例中，为了使读者更好地理解本申请而提出了许多技术细节。但是，即使没有这些技术细节和基于以下各实施例的种种变化和修改，也可以实现本申请所要求保护的技术方案。
参看图1，本申请实施例提供了一种丘脑底核深部脑刺激术电极开机触点选择方法，包括以下步骤：
步骤S1、采集待测患者的不同脑区的MERs信号。
步骤S2、从MERs信号中分别提取时序特征与幅序特征。
步骤S3、将时序特征与幅序特征融合，得到MERs信号的幅频感知融合特征图。
步骤S4、将幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置。
步骤S5、对STN的边界以及永久治疗电极的置入位置进行分析，确定最优刺激触点。
针对术后最佳触点选择过程需要反复试验且耗时费力的难点，本申请提出了一种基于幅频感知深度融合网络的STN-DBS电极最优触点的选择方法。首先，提出将MERs信号的时间序列特性和生成的GASF特征相结合得到幅频感知融合特征，然后提出幅频感知深度融合网络对不同频率和幅值的微电极记录分级信号进行分类，可以帮助临床医生有效地识别STN边界和微电极植入的最佳位置，以实现DBS永久电极的最佳触点的选择。
具体的，本申请提出的基于幅频感知深度融合网络的STN-DBS电极最优触点选择方法，该方法首先识别STN的边界用于指导永久治疗电极的置入，然后确定微电极植入最佳位置的聚集区，达到精确选择术后最优开机触点的目的。具体的，一方面，本申请利用二维时间序列图像和GASF图像分别表征MERs时间序列信号的高频和高幅特征。该方法可以将时间序列信号分级识别任务转换为利用AI技术进行二维融合的视觉图像特征分类任务；另一方面，本申请提出幅频感知深度融合网络，根据不同的频率和幅值对分级的MERs信号进行自动识别，可以帮助临床医生自动的识别STN边界和微电极植入的最佳位置。最后，本申请还提出了一种STN-DBS手术中、术后自动决策的方法。该方法准确地识别了STN边界用于永久治疗电极置入位置的确定和最优开机触点选择。实验结果表明，神经外科医生和本文方法选择了相同的最佳开机触点。此外，该方法在STN和微电极植入最佳位置的识别方面的性能优于目前最先进的方法。
本申请提出的基于幅频感知深度融合网络的STN-DBS电极最优触点的选择方法的总流程图如图2所示，针对非连续的高频、高幅信号和峰值不明显的MERs信号识别的难题以及术后最佳触点选择过程需要反复试验且耗时费力的难题，选择方法的整个过程被总结为两个阶段：基于幅频感知深度融合网络的MERs信号的自动识别和最优刺激触点的选择。STN-DBS术中和术后难点主要包括STN边界定位、微电极的最佳植入位置识别以及永久治疗电极最优触点的选择，本 申请为STN-DBS术中和术后的这些难点提供了智能决策方案。首先，通过将CWT提取的时序特征与GASF特征融合得到MERs信号的幅频感知融合特征图。然后利用CBAM集成CNN模型用于识别幅频感知的融合特征图，用于识别STN的边界和微电极植入最佳位置。在STN边界被确定之后，DBS永久电极的置入原则是电极头尽量触及STN的下边界且不会影响黑质，同时要避免对患者产生不良反应。最后，对永久治疗电极置入的位置和微电极植入的最佳位置的分布综合分析可以确定最优的刺激触点。
步骤S1是采集MERs信号的过程。MERs信号是通过一种电生理导航系统(NeuroNav 223/12，Alpha Omega工程，以色列)采集。神经外科医生首先利用MRI影像初步的确定靶区位置，并结合CT扫描技术规划微电极植入轨迹，然后根据预设的靶区位置和规划的电极的轨迹将微电极植入STN核团。一般情况下，MERs的测量方法如下：(1)信号的采集通过植入一个或两个钨微电极；(2)微电极(带针芯的套管)被插入到预设的靶区上方约10mm处；(3)微电极进一步以大约1mm步长移动直到插入距离预设靶区约5mm，每步测量一个MERs信号；(4)微电极再以大约0.5mm步长进行推进直至到靶区之下约4.0～5.0mm，每一步测量一个MERs信号。微电极的运动可能会产生噪声干扰，临床医生通常需要在每个位置多等待一段时间使MERs信号达到稳定，因此临床在微电极的每个位置收集的MERs信号时长是不一样的，通常采集1～5s的信号。考虑到不同长度的MERs数据可能会影响信号处理的实验结果，为排除噪声干扰并保持MERs信号的稳定性，根据专业神经外科医生的建议，每个位置选取噪声最小的1s时长的MERs信号片段。原始的MERs信号由专业神经外科医生使用Mapfile Converter软件(版权Alpha Omega 2018，版本5.1.11.0)进行预处理，然后利用MATLAB软件(Mathworks，Inc.，Natick，MA)通过非重叠的1秒时长的滑动窗对MERs进行裁剪并标记。所有信号采样频率为24kHz，放大倍数为1000，在MATLAB中使用0.5kHz至5kHz的6极Butterworth数字带通滤波器和60Hz的陷波滤波器进行数字滤波。所有原始MERs信号通过适当的抗混叠滤波降采样 至12kHz，可以减少数据采样点，提高信号识别的效率。微电极通常贯穿大脑的以下区域，主要包括Rt、IC、ZI、H2、STN和SN。由于不同的脑区具有不同的电生理信号，因此通过采集的不同脑区的MERs信号特征的差异可以有效的识别STN边界和微电极植入的最佳位置。根据采集的MERs信号识别的STN边界位置和微电极探测的植入路径，神经外科医生可以确定永久治疗电极的植入轨迹，然后将所有的微电极移除，最后的四触点的永久治疗电极置入到STN。在术后阶段，临床医生根据临床经验选择最优触点以改善帕金森患者的运动症状。本研究得到了当地医院人体研究伦理委员会的批准。在对所涉及的程序进行充分的解释后，获得了所有参与者的书面知情同意。
在一些实施例中，步骤S2中从MERs信号中分别提取时序特征与幅序特征，包括：采用连续小波变换方法，从MERs信号中提取时序特征；以及采用GASF方法将MERs信号转换为GASF视觉图像，并基于GASF视觉图像，提取幅序特征。
具体的，在STN-DBS手术中，微电极植入的轨迹从前额叶运动皮层开始一直到达黑质。为了避免电刺激黑质，精确的判定治疗电极的置入位置在手术过程中至关重要。治疗电极置入STN之后，可根据术中微电极植入的最佳位置的分布选择最佳触点。因此，STN-DBS的挑战在于识别STN的边界以确定治疗电极的置入位置，以及识别微电极植入的最佳位置选择术后的最佳触点。由于不同脑组织的MERs信号具有不同的特征，因此可以根据MERs的特征来识别STN的边界和微电极植入的最佳位置。对于微电极植入的最佳位置的识别，经验丰富的临床医生通常根据峰值、频率和振幅特征将MERs信号分5级标记，位于微电极植入的最佳位置的MERs属于最高两个等级，其余为非最佳位置的MERs。因此，STN的边界和微电极植入的最佳位置的识别可以定义为一个典型的三分类问题，即MERs信号来自非STN、STN中的最佳位置和STN中的非最佳位置。
由于MERs信号具有随机性、多分量性和非平稳性等复杂特征，将一维的MERs信号转化为二维图像，可以克服MERs复杂一维特征分析的难题。MERs时间序列图片转换最常用的两种特征提取方法是时序特征提取和幅序特征提取。 连续小波变换作为一种时频域变换方法，具有分析非平稳信号的优势，可以有效地提取时间序列特征。
在一些实施例中，给定一个时间序列信号，连续小波变换CWT表示为：

公式(1)和公式(2)中，ψ为小波基函数，a为尺度系数将ψ(t)扩展或缩小到ψ(t/a)，a≠0，b为小波位置的时移系数。
在一些实施例中，采用GASF方法将所述MERs信号转换为GASF视觉图像，包括：将MERs时间序列归一化至区间[-1,1]，得到归一化后的MERs数据；MERs时间序列表示为Y＝{y1，y2，...，yn}；将归一化后的MERs数据转换为极坐标；基于极坐标，得到GASF矩阵；基于GASF矩阵，生成GASF视觉图像。
复Morlet小波作为一种应用广泛的小波基函数，在CWT中对时间序列特征提取具有最优的时频分辨率和最优的聚焦性能，因此，本申请选择复Morlet小波(cmor3-3)作为小波基函数。由于尺度序列越长，频率分辨率越高，而尺度序列过长可能会造成时间分辨率的损失，因此，尺度序列长度设为256。由于信号的采样频率为12kHz，则采样时间为，总的时间序列长度为1s，时移系数序列可以通过采样时间沿时间轴得到。通过改变时移系数计算小波系数，得到最终的时频特征图。利用典型的时间序列特征提取方法CWT，可以将位于STN的最佳位置、STN的非最佳位置和非STN的MERs信号生成3维的时频图像。
为了获得MERs信号的幅序特征，利用GASF方法将MERs信号转换为GASF视觉图像。已知MERs时间序列，Y被归一化到区间[-1,1]。然后将归一化后的MERs数据转换为极坐标表示，如公式(3)所示：
通过检验三角和，可以很容易地识别出不同时间区间内的时间相关性，则GASF矩阵如公式(4)所示：
其中，角度φ和半径r利用归一化的MERs的值和相应的时间戳t来表示，其中N是正则化跨度的常数因子。
然后将GASF矩阵生成尺寸为100×100×3的GASF图像。来自STN的的最佳位置、STN的非最佳位置和非STN的三类MERs信号生的GASF图像具有纹理和条纹特征差异。
在一些实施例中，幅频感知融合网络采用CBAM集成的卷积神经网络模型为分类器，对幅频感知融合特征图进行识别，得到STN边界以及永久治疗电极的置入位置。
为了准确地识别STN的边界和微电极植入的最佳位置，将CWT得到的时序特征图与GASF方法得到的幅序特征图相融合得到幅频感知融合特征，得到的幅频感知融合特征图的维度为100×100×6。由于注意力机制可以重点提取MERs信号的幅频感知融合特征中利于分类的特征信息，抑制不相关的神经活动特征，因此本申请提出CBAM集成的CNN模型为分类器用于识别幅频感知融合特征图，可以精确的识别STN边界和微电极植入的最佳位置。
在一些实施例中，幅频感知深度融合网络采用七个阶段执行全分辨率的卷积，每个阶段包括一个卷积层模块和一个CBAM模块；最后一个阶段包括一个全连接层和一个Softmax函数，用来获得幅频感知深度融合网络的输出。
具体的，为了在分类任务中能够提取更多丰富的信息同时避免池化造成的信息丢失，该幅频感知深度融合网络采用七个阶段执行全分辨率的卷积。同时，考虑到输入图像尺寸设置为100×100×3，经过四个阶段卷积之后神经元的接受域就足以覆盖输入，因此在最后三个阶段使用池化层来获得用于分类任务的高级特征。如图3所示，该幅频感知深度融合网络由七个阶段组成，每个阶段包含一个卷积层模块和一个CBAM模块。最后一个阶段是一个全连接层和一个Softmax函数用来获得整个网络的输出。Maxpooling层分别跟在第四、第五和第六阶段后面。每个卷积层之后是BN和ReLU。
如图3所示，在一些实施例中，CBAM模块包括通道注意力模块和空间注意力模块；给定一个中间特征映射X作为输入，整个CBAM过程表示为：
其中，MC为通道注意力模块，MS为空间注意力模块，表示逐个元素相乘，X′表示通过通道注意力模块后获得的特征，X″为CBAM的最终输出。
对于通道注意力模块，输入特征映射同时经过平均池化和最大池化，然后通过多层感知机。多层感知机包含一个共享的MLP，然后对MLP的输出特性进行元素求和，再通过sigmoid激活函数进行输出。则通道注意力模块的计算公式如下所示：
MC(X)＝σ(MLP(fAvgPool(X))+MLP(fMaxPool(X)))     (6)
其中，MC为通道注意力模块，fAvgPool和fMaxPool分别为平均池化和最大池化；MLP表示多层感知机，σ为sigmoid函数。
然后将通道注意力模块的输出特征图作为空间注意力模块的输入特征图。首先，通过最大池化和平均池化处理输入特征，在通道维上进行压缩采样，然 后经过一个卷积层，得到二维空间注意力特征图；空间注意力模块的计算公式如下所示：
MS(X′)＝σ(Fcov([fAvgPool(X)；fMaxPool(X)]))     (7)
其中，MS为空间注意力模块，Fcov表示卷积运算，卷积核大小为7×7。
CBAM模块能够有效的学习强调或抑制的特征，并有效地细化中间特征。在本申请提出的幅频感知深度融合网络中，第1～7阶段卷积层的通道数分别为16、32、64、128、128、128、128、128。卷积核大小为3×3，最大池化层的核大小为2×2。模型网络初始化采用“random_uniform”方法，采用Adam算法进行网络优化。学习率和批量大小分别设置为1e-2和32。在训练阶段，迭代次数设置为50。
根据上述模型分析，基于幅频感知深度融合网络的MERs信号识别算法流程如表1所示：
表1基于幅频感知融合网络算法的实现过程

具体的算法描述如下：
如表1所示，DCWT＝{a1,a2,…,a100}和DGASF＝{b1,b2,…,b100}分别表示CWT和GASF方法生成的特征图组成的样本集，其中，a1和b1分别表示第一个患者的CWT和GASF特征图集合，以此类推。首先将DCWT和DGASF样本集对应样本进行拼接得到融合特征图样本集X，样本集X表示100人的MERs数据生成的两种融合特征图组成的集合。第二，将100人样本集拆分为80人用于训练，20人用于测试，则可以得到训练集Xtr＝{c1,c2,…,c80}和测试集Xtest＝{c81,c82,…,c100}。第三，将Xtr以8人为一组，分为10个块Xtr＝{x1,x2,…,x10}，其中表示以8人为一组组成的特征图集合。第四，采用十折交叉方法将训练集中的一块数据用作验证集Xval，则训练集中其余数据更新为新的训练集，将得到的新的训练集和验证集输入到CBAM增强网络模型中，用于训练模型参数。第五，在测试过程，将测试集输入到训练好的模型中，得到预测的标签值ypre，然后根据测试集标签ytest计算预测结果与实际结果的混淆矩阵，可以得到算法的分类准确率、精确率等评价结果，用来评价模型性能。
在一些实施例中，步骤S5中对STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点，包括：对STN的边界以及永久治疗电极的置入位置进行分析，得到永久治疗电极的置入策略；基于永久治疗电极的置入策略，确定最优刺激触点；永久治疗电极的置入策略为永久治疗电极的顶端靠近STN的下边界；最优刺激触点的选择的原则是覆盖永久治疗电极植入的最佳位置的聚集区。
最优刺激触点的选择包括永久治疗电极的置入策略以及术后最优开机触点的选择策略。
对于永久治疗电极的置入策略，永久治疗电极置入位置会严重影响最优触点的选择，其原因是永久治疗电极置入过深，若超过STN的下边界，触及黑质会对患者造成不良后果，严重的可能诱发各种并发症，若电极没有完全置入STN中，则对靶区术后开机程控的覆盖面不够，不能完全改善患者的运动状态。永久治疗电极置入之前，首先需要利用微电极多次探测以找到STN的边界和微电极植入的最佳位置，第三节中已经介绍了根据MERs信号识别STN的边界和电极植入的最佳位置的主要过程。由于永久治疗电极置入需要受多个条件的限制，因此如何根据STN的边界和微电极植入最佳位置以及其他的限制条件制定一个合理的永久治疗电极的置入策略也是一个难点。
永久治疗电极置入受多个条件限制，主要包括STN的尺寸、永久治疗电极的尺寸、触点电刺激的覆盖区域，具体描述如下：
(1)STN的尺寸。依据MERs信号特征可以得到STN边界范围为-1到4.5点区间范围，则得到实际的STN的尺寸为5.5mm，而且STN的下边界也有明显差异，因此永久治疗电极的置入要严格参照实际识别的STN区域。
(2)永久治疗电极的尺寸。永久治疗电极的长度约为7.5mm，由四个1.5mm长的触点组成，间隔0.5mm。每个触点都可以外加电刺激，对靶区进行程控，而且程控的区域也是有限的。
(3)触点电刺激的覆盖区域。选择的最优电极触点外加电刺激要尽量覆盖STN，尤其是感觉运动区域，同时又需要避免刺激SN，避免术后程控对患者带来副作用。
因此，综合分析STN的尺寸、永久治疗电极的尺寸、触点电刺激的覆盖区域这三个限制因素，本申请将DBS永久治疗电极置入的最佳位置总结为电极的顶端要靠近STN下边界。
对于术后最优开机触点的选择策略，永久治疗电极置入STN之后，在胸前植入脉冲发生器，则STN-DBS手术完成。患者STN-DBS手术之后，需要经过开机程控即在治疗电极的最优触点上加入电刺激以达到改善患者运动状态的目的。最优开机触点应尽可能的电刺激感觉运动区，才能达到最佳治疗效果。由于微电极植入的最佳位置的聚集区可视为STN的感觉运动区，因此根据微电极植入的最佳位置的分布可以确定最佳的刺激触点，同时还应该注意避免电刺激影响SN。通过MERs信号可以有效的识别出微电极植入的最佳位置，最优刺激触点的选择的原则是覆盖微电极植入的最佳位置的聚集区。微电极植入的最佳位置的聚集区主要是位于电极深度为1～2.5mm区间内，则该区间覆盖的电极触点为标号2触点，因此术后的最佳刺激触点选择是以微电极植入的最佳位置的聚集区为重要依据。
通过本申请提供的丘脑底核深部脑刺激术电极开机触点选择方法进行实验验证，结果证实本申请的微电极植入的最佳位置的识别准确率为97.47％，能够满足临床准确率要求。此外，本申请能够实现从20个测试患者中精准识别出18个患者的最优刺激触点，因此最优刺激触点准确率Accuracy＝18/(20*100％＝)90％。在临床实践中，临床医生通常选择最优和次最优电刺激触点作为最终候选触点。结合临床实践，将次最优触点考虑到最终结果评价当中，则本申请能有效的识别出20个测试患者的最优触点，准确率可以达到100％，满足精准手术的要求。因此，本申请的选择与临床医生的选择一致，可以为神经外科医生提供有用的指导。
参看图4，本申请实施例还提供了一种丘脑底核深部脑刺激术电极开机触点选择系统，包括：依次连接的MERs信号采集模块101、特征提取模块102、特征融合模块103、识别模块104以及分析模块105；MERs信号采集模块101用于采集待测患者的不同脑区的MERs信号；特征提取模块102用于从所述MERs信号中分别提取时序特征与幅序特征；特征融合模块103用于将所述时序特征与所述幅序特征融合，得到MERs信号的幅频感知融合特征图；识别模块104用于将所述幅频感 知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置；分析模块105用于对所述STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点。
参看图5，本申请另一实施例提供了一种电子设备，包括：至少一个处理器110；以及，与至少一个处理器通信连接的存储器111；其中，存储器111存储有可被至少一个处理器110执行的指令，指令被至少一个处理器110执行，以使至少一个处理器110能够执行上述任一方法实施例。
其中，存储器111和处理器110采用总线方式连接，总线可以包括任意数量的互联的总线和桥，总线将一个或多个处理器110和存储器111的各种电路连接在一起。总线还可以将诸如外围设备、稳压器和功率管理电路等之类的各种其他电路连接在一起，这些都是本领域所公知的，因此，本文不再对其进行进一步描述。总线接口在总线和收发机之间提供接口。收发机可以是一个元件，也可以是多个元件，比如多个接收器和发送器，提供用于在传输介质上与各种其他装置通信的单元。经处理器110处理的数据通过天线在无线介质上进行传输，进一步，天线还接收数据并将数据传送给处理器110。
处理器110负责管理总线和通常的处理，还可以提供各种功能，包括定时，外围接口，电压调节、电源管理以及其他控制功能。而存储器111可以被用于存储处理器110在执行操作时所使用的数据。
本申请另一实施例涉及一种计算机可读存储介质，存储有计算机程序。计算机程序被处理器执行时实现上述方法实施例。
由以上技术方案，本申请实施例提供一种丘脑底核深部脑刺激术电极开机触点选择方法及系统，该方法包括以下步骤：首先，采集待测患者的不同脑区的MERs信号；从MERs信号中分别提取时序特征与幅序特征；将时序特征与幅序特征融合，得到MERs信号的幅频感知融合特征图；将幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN 的边界以及永久治疗电极的置入位置；对STN的边界以及永久治疗电极的置入位置进行分析，确定最优刺激触点。
本申请提供的丘脑底核深部脑刺激术电极开机触点选择方法，是基于幅频感知深度融合网络的STN-DBS电极最优触点的选择方法。该方法将幅频感知融合特征和CBAM集成CNN模型用于STN边界和微电极植入的最佳位置的识别，根据STN边界指导STN-DBS永久治疗电极的置入，依据微电极植入的最佳位置分布选择术后最佳的开机触点。实验结果表明，本申请STN识别的平均准确率为96.78％，精确率为97.77％，灵敏度为95.26％，特异度为98.07％；微电极植入的最佳位置识别的平均准确率为98.24％，精确率为98.53％，灵敏度为96.05％，特异度为99.31％，所有结果均优于目前的STN和微电极植入的最佳位置识别算法。所提出的方法选择的最优触点与神经外科医生的成功决策结果一致，这表明本申请所提出的方法在提高手术效率的同时减少对临床医生经验的依赖有很大的潜力。
即，本领域技术人员可以理解，实现上述实施例方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成，该程序存储在一个存储介质中，包括若干指令用以使得一个设备(可以是单片机，芯片等)或处理器(processor)执行本申请各个实施例上述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
本领域的普通技术人员可以理解，上述各实施方式是实现本申请的具体实施例，而在实际应用中，可以在形式上和细节上对其作各种改变，而不偏离本申请的精神和范围。任何本领域技术人员，在不脱离本申请的精神和范围内，均可作各自更动与修改，因此本申请的保护范围应当以权利要求限定的范围为准。

Claims (10)

  1. 一种丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，包括：
    采集待测患者的不同脑区的MERs信号；
    从所述MERs信号中分别提取时序特征与幅序特征；
    将所述时序特征与所述幅序特征融合，得到MERs信号的幅频感知融合特征图；
    将所述幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置；
    对所述STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点。
  2. 根据权利要求1所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，从所述MERs信号中分别提取时序特征与幅序特征，包括：
    采用连续小波变换方法，从所述MERs信号中提取时序特征；以及采用GASF方法将所述MERs信号转换为GASF视觉图像，并基于所述GASF视觉图像，提取幅序特征。
  3. 根据权利要求2所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，所述连续小波变换表示为：

    公式(1)和公式(2)中，ψ为小波基函数，a为尺度系数将ψ(t)扩展或缩小到ψ(t/a)，a≠0，b为小波位置的时移系数。
  4. 根据权利要求2所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，采用GASF方法将所述MERs信号转换为GASF视觉图像，包括：
    将MERs时间序列归一化至区间[-1,1]，得到归一化后的MERs数据；所述 MERs时间序列表示为Y＝{y1，y2，…，yn}；
    将所述归一化后的MERs数据转换为极坐标；
    基于所述极坐标，得到GASF矩阵；
    基于所述GASF矩阵，生成GASF视觉图像；
    所述极坐标如公式(3)所示：
    所述GASF矩阵如公式(4)所示：
    其中，角度φ和半径r利用归一化的MERs的值和相应的时间戳t来表示，其中N是正则化跨度的常数因子。
  5. 根据权利要求1所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，所述幅频感知融合网络采用CBAM集成的卷积神经网络模型为分类器，对所述幅频感知融合特征图进行识别，得到STN边界以及永久治疗电极的置入位置。
  6. 根据权利要求5所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，所述幅频感知深度融合网络采用七个阶段执行全分辨率的卷积，每个阶段包括一个卷积层模块和一个CBAM模块；最后一个阶段包括一个全连接层和一个Softmax函数，用来获得所述幅频感知深度融合网络的输出。
  7. 根据权利要求6所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，所述CBAM模块包括通道注意力模块和空间注意力模块；
    给定一个中间特征映射X作为输入，整个CBAM过程表示为:
    其中，Mc为通道注意力模块，MS为空间注意力模块，表示逐个元素相乘，X′表示通过通道注意力模块后获得的特征，X″为CBAM的最终输出。
  8. 根据权利要求7所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，所述通道注意力模块的计算公式如下所示：
    MC(X)＝σ(MLP(fAvgPool(X))+MLP(fMaxPool(X)))   (6)
    其中，MC为通道注意力模块，fAvgPool和fMaxPool分别为平均池化和最大池化；MLP表示多层感知机，σ为sigmoid函数；
    所述空间注意力模块的计算公式如下所示：
    MS(X′)＝σ(Fcov([fAvgPool(X)；fMaxPool(X)]))   (7)
    其中，MS为空间注意力模块，Fcov表示卷积运算。
  9. 根据权利要求1所述的丘脑底核深部脑刺激术电极开机触点选择方法，其特征在于，对所述STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点，包括：
    对所述STN的边界以及所述永久治疗电极的置入位置进行分析，得到永久治疗电极的置入策略；基于所述永久治疗电极的置入策略，确定最优刺激触点；
    所述永久治疗电极的置入策略为永久治疗电极的顶端靠近STN的下边界；
    所述最优刺激触点的选择的原则是覆盖永久治疗电极植入的最佳位置的聚集区。
  10. 一种丘脑底核深部脑刺激术电极开机触点选择系统，其特征在于，包括：依次连接的MERs信号采集模块、特征提取模块、特征融合模块、识别模块以及分析模块；
    所述MERs信号采集模块用于采集待测患者的不同脑区的MERs信号；
    所述特征提取模块用于从所述MERs信号中分别提取时序特征与幅序特征；
    所述特征融合模块用于将所述时序特征与所述幅序特征融合，得到MERs信号的幅频感知融合特征图；
    所述识别模块用于将所述幅频感知融合特征图输入至幅频感知融合网络，对STN的边界和微电极植入最佳位置进行识别，确定STN的边界以及永久治疗电极的置入位置；
    所述分析模块用于对所述STN的边界以及所述永久治疗电极的置入位置进行分析，确定最优刺激触点。
PCT/CN2023/137693 2023-06-21 2023-12-09 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统 Ceased WO2024259913A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202310743390.0 2023-06-21
CN202310743390.0A CN116943027A (zh) 2023-06-21 2023-06-21 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统

Publications (1)

Publication Number Publication Date
WO2024259913A1 true WO2024259913A1 (zh) 2024-12-26

Family

ID=88445315

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/137693 Ceased WO2024259913A1 (zh) 2023-06-21 2023-12-09 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统

Country Status (2)

Country Link
CN (1) CN116943027A (zh)
WO (1) WO2024259913A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116943027A (zh) * 2023-06-21 2023-10-27 中国科学院深圳先进技术研究院 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100100152A1 (en) * 2007-03-02 2010-04-22 Koninklijke Philips Electronics N.V. Electrode system for deep brain stimulation
WO2018033732A1 (en) * 2016-08-16 2018-02-22 Oxford University Innovation Limited Checking the location of implantation of a deep brain stimulation (dbs) lead
CN110337312A (zh) * 2017-02-28 2019-10-15 博医来股份公司 基于刺激场建模的最佳脑深部刺激电极的选择与放置
CN113797440A (zh) * 2021-09-27 2021-12-17 首都医科大学附属北京天坛医院 基于影像和电生理实时定位的脑深部电极自动植入系统
WO2022211976A1 (en) * 2021-03-31 2022-10-06 The Regents Of The University Of Michigan System and method to select among trajectories for therapeutic stimulation of a target volume region within the brain
CN116943027A (zh) * 2023-06-21 2023-10-27 中国科学院深圳先进技术研究院 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100100152A1 (en) * 2007-03-02 2010-04-22 Koninklijke Philips Electronics N.V. Electrode system for deep brain stimulation
WO2018033732A1 (en) * 2016-08-16 2018-02-22 Oxford University Innovation Limited Checking the location of implantation of a deep brain stimulation (dbs) lead
CN110337312A (zh) * 2017-02-28 2019-10-15 博医来股份公司 基于刺激场建模的最佳脑深部刺激电极的选择与放置
WO2022211976A1 (en) * 2021-03-31 2022-10-06 The Regents Of The University Of Michigan System and method to select among trajectories for therapeutic stimulation of a target volume region within the brain
CN113797440A (zh) * 2021-09-27 2021-12-17 首都医科大学附属北京天坛医院 基于影像和电生理实时定位的脑深部电极自动植入系统
CN116943027A (zh) * 2023-06-21 2023-10-27 中国科学院深圳先进技术研究院 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
XIAO, LINXIA ET AL.: "Amplitude-frequency-aware deep fusion network for optimal contact selection on STN-DBS electrodes", SCIENCE CHINA, vol. 65, 10 March 2022 (2022-03-10), XP037730643, DOI: 10.1007/s11432-021-3392-1 *

Also Published As

Publication number Publication date
CN116943027A (zh) 2023-10-27

Similar Documents

Publication Publication Date Title
US11439826B2 (en) Systems and methods for clinical decision making for a patient receiving a neuromodulation therapy based on deep learning
US7177674B2 (en) Patient-specific parameter selection for neurological event detection
US9603526B2 (en) Systems and methods for compound motor action potential monitoring with neuromodulation of the pelvis and other body regions
WO2024098449A1 (zh) 基于深度学习和电生理信号的癫痫病灶区定位系统及方法
US20240366315A1 (en) Brain navigation lead positioning and methods thereof
CN119386383A (zh) 一种多功能磁刺激治疗系统及其应用方法
CN116492600B (zh) 基于个体化时-空靶点的调控装置、设备及存储介质
CN116340735B (zh) 一种基于长视野特征融合的棘波检测方法及系统
CN114712712B (zh) 刺激电极导线的成像识别方法及相关装置
CN116849609A (zh) 一种致痫区自动定位方法、系统、设备及存储介质
CN119565030A (zh) 一种用于认知功能障碍的神经调控系统
CN120242328A (zh) 一种闭环神经调控方法、设备、介质和程序产品
WO2023221056A1 (zh) 医疗调控系统及其使用方法和可读存储介质
WO2024259913A1 (zh) 一种丘脑底核深部脑刺激术电极开机触点选择方法及系统
WO2025001908A1 (zh) 脉冲发生器、刺激器、存储介质及程序产品
Adochiei et al. Determination of cognitive-emotional characteristics through the method of individual asymmetries of power spectral density of EEG waves
Maged et al. Lightweight deep learning model for automated STN localization using MER in Parkinson’s disease
CN107822628B (zh) 一种癫痫脑部病灶区自动定位装置及系统
Xiao et al. Amplitude-frequency-aware deep fusion network for optimal contact selection on STN-DBS electrodes
CN106725463A (zh) 应用皮层脑电信号对大脑皮层手部功能区定位的方法和系统
US20250201427A1 (en) Machine learning based automated neurostimulation programming
Li et al. Automatic detection of scalp high-frequency oscillations based on deep learning
CN116680571A (zh) 参数配置装置、医疗系统、存储介质及程序产品
US11865327B1 (en) System and method for identifying organelles and tissue types of the brain using representative frequency signatures
US20240379233A1 (en) Systems and Methods for Generating Reference Limits for Use in Diagnostic and Monitoring Systems

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23942179

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE