Multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm

Abstract

The invention provides a multi-group brain electrode three-dimensional implantation path planning system based on a genetic algorithm, and the system comprises a module M1 which collects the imaging data of a brain and the depth information of electrodes, and generates a three-dimensional brain model; the module M2 is used for generating candidate paths based on the three-dimensional brain model; constructing an objective function, analyzing the candidate path, and selecting a corresponding path based on a quality analysis result of the candidate path; and the module M3 is used for optimizing the corresponding path based on the real-time image and the target function to obtain an optimal path. According to the method, the problems that the distance between the electrodes is difficult to balance, the key brain structure is difficult to damage and the implantation precision is difficult to guarantee during multi-electrode implantation in a traditional artificial planning method are solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of neural engineering, and particularly relates to a multi-group brain electrode three-dimensional implantation path planning system based on a genetic algorithm. BACKGROUND

[0002] Currently, the path planning of clinical brain deep electrode implantation highly depends on the personal experience of neurosurgeons. The doctor needs to manually design the path of a single electrode on a two-dimensional image, and repeatedly adjust to avoid key blood vessels and functional areas. The process is tedious and time-consuming, and the result is greatly affected by subjective factors. Although there are auxiliary software that can semi-automatically optimize the path of a single electrode based on three-dimensional images, there are obvious limitations when dealing with multi-electrode collaborative implantation: existing tools usually adopt a "sequential planning" strategy, that is, after planning one, the next one is processed. This causes the subsequent electrodes to be forced to adapt to the already determined path, and it is impossible to achieve global collaborative optimization of multi-electrode trajectories, which can easily cause problems such as electrode path congestion and intersection, insufficient avoidance structure, or undesirable implantation angle, directly affecting the safety of surgery and the accuracy of efficacy evaluation.

[0003] The core defect of the current method is the lack of efficient and automated global collaborative planning capability for multiple electrodes. The specific performance is that: existing tools are difficult to balance the complex constraints in the multi-electrode scenario; the search space of the multi-electrode combined path is huge, and traditional optimization algorithms cannot find high-quality solutions within a clinically acceptable time.

[0004] Patent document CN117481799A discloses a path planning method, device and equipment for brain electrode implantation surgery. The disadvantages of this scheme are that the implantation path planning algorithm is simple, and the influence of multiple factors in electrode implantation and how to adjust the implantation strategy online are not considered.

[0005] Patent document CN115445082A discloses an automatic brain deep electrode implantation system. The scheme includes: a segmentation module for acquiring multi-modal images, segmenting the multi-modal images, and determining brain blood vessels and target nuclei; a path planning module for determining the implantation position of the electrode implantation target point based on the brain blood vessels and the target nuclei, and generating a target surgery path; a collection module for controlling the movement of the electrode based on the target surgery path during the surgery, and collecting the neural signals of the target object based on the electrode; a judgment module for displaying the real-time position of the electrode on the target surgery path, and combining the neural signals and the real-time position to determine whether the electrode enters the target nucleus region; and a sending module for generating a prompt information when the electrode enters the target nucleus region. This scheme cannot consider other influencing factors and calibrate the implantation strategy online.

[0006] This problem needs to be solved urgently. SUMMARY

[0007] Aiming at the defects in the prior art, the present application aims to provide a multi-group brain electrode three-dimensional implantation path planning system based on a genetic algorithm.

[0008] According to the present application, a multi-group brain electrode three-dimensional implantation path planning system based on a genetic algorithm is provided, comprising:

[0009] Module M1: collect imaging data of the brain and depth information of the electrodes, and generate a three-dimensional brain model;

[0010] Module M2: generate a candidate path based on the three-dimensional brain model; construct an objective function, analyze the candidate path, and select a corresponding path based on the quality analysis result of the candidate path;

[0011] Module M3: optimize the corresponding path based on real-time images and the objective function to obtain an optimal path.

[0012] Preferably, in the module M2, the mathematical expression of the objective function is:

[0013]

[0014] wherein J(P) represents the objective function; P represents the position information of the electrodes; ∫ represents integration; represents the union set of all electrode coverage areas; w(x) represents the information density at position x.

[0015] Preferably, in the module M2, the mathematical expression of the objective function is:

[0016]

[0017] wherein J(P) represents the objective function; represents the union set of all electrode coverage areas; w(x) represents the information density at position x; P represents the position information of the electrodes; ∫ represents integration; λ1, λ2, λ3 and λ4 represent the coefficients of C vessel , C depth , C spacing and C length , respectively; C vessel represents a blood vessel avoidance constraint; C depth represents an implantation depth limit; C spacing represents an electrode spacing limit; C length represents a path length limit.

[0018] Preferably, in the module M2, the candidate path is analyzed to obtain a quality analysis result, which comprises:

[0019] Module M2.1: adopt a two-point crossover strategy to select a candidate path to obtain a preliminary selected path;

[0020] The double-point intersection strategy, i.e., intersecting two implantation paths from the surface implantation site and the implantation direction, and then obtaining a new path, i.e., a candidate path; the expression of the starting point of the candidate path is:

[0021] p new =(1-λ)p A +λp B

[0022] wherein p new represents the starting point of the candidate path; λ represents the coefficient of the double-point intersection strategy; p A and p B respectively represent the two existing implantation paths; when the two implantation paths are intersected, spherical linear interpolation is performed along the implantation direction;

[0023] Module M2.2: judge whether the objective function metric of the primary selected path is ranked according to the current objective function, and the result is yes, then retain the individual whose objective function score is higher than the preset threshold, and perform the next round of iteration until the preset maximum iteration number is reached; the result is no, then not retained;

[0024] Preferably, the mathematical expression of C vessel (P) is:

[0025]

[0026] wherein N represents the number of electrodes to be implanted; i represents the i-th electrode; d min represents the minimum distance between the electrode and the blood vessel; l represents a voxel occupied by the i-th electrode; represents the voxel set occupied by the i-th electrode; v represents a voxel in the blood vessel voxel set; represents the entire voxel set occupied by all blood vessels.

[0027] Preferably, the mathematical expression of C depth (P) is:

[0028]

[0029] wherein h i represents the implantation depth of the electrode, h max represents the maximum implantation depth; h min represents the minimum implantation depth.

[0030] Preferably, the mathematical expression of C spacing (P) is:

[0031]

[0032] wherein s minIndicates the minimum permissible spacing between electrodes; p i This represents the position information of the i-th electrode; p j This represents the position information of the j-th electrode.

[0033] Preferably, C length The mathematical expression for (P) is:

[0034]

[0035] Among them, L i L represents the total path length from the robot's zero position to the planned implantation site; max Indicates the maximum length.

[0036] Preferably, in module M3, the objective function metric of the initially selected path is iterated, and it is determined whether the accuracy improvement magnitude ΔJ < ∈, or the maximum number of iterations is greater than T. max If the result is yes, then stop the iteration and output the optimal path as the quality analysis result; if the result is no, then do not process it.

[0037] Wherein, ∈ represents the minimum objective function lifting threshold; T max This represents the maximum number of iterations required for the genetic algorithm.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. This invention solves the problems of traditional manual planning methods in balancing the distance between electrodes, avoiding damage to key brain structures, and ensuring implantation accuracy when implanting multiple electrodes by executing the process of the three-dimensional implantation path planning method based on the genetic algorithm.

[0040] 2. This invention successfully implants multiple sets of electrodes in complex brain environments, significantly improving the efficiency and success rate of implantation in limited space, achieving hemorrhage sites all smaller than 150 micrometers, and an overall success rate of over 80%.

[0041] 3. This invention comprehensively considers factors such as three-dimensional vascular distribution, implantation depth, and information acquisition gain, and can achieve the optimal spatial distribution of multiple electrode sets, which can significantly improve the safety and accuracy of multiple electrode implantation and meet the needs of high-throughput implantation. Attached Figure Description

[0042] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0043] Figure 1 This is a schematic diagram of multiple sets of brain electrodes implanted according to the present invention;

[0044] Figure 2 This is a flowchart provided for the present invention;

[0045] Figure 3 This is a schematic diagram illustrating the optimization results of the implantation path based on multiple constraints provided by the present invention. Detailed Implementation

[0046] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0047] The present invention also provides a three-dimensional implantation path planning system for multiple sets of brain electrodes based on a genetic algorithm. The three-dimensional implantation path planning system for multiple sets of brain electrodes based on a genetic algorithm can be implemented by executing the process steps of the three-dimensional implantation path planning method for multiple sets of brain electrodes based on a genetic algorithm. That is, those skilled in the art can understand the three-dimensional implantation path planning method for multiple sets of brain electrodes based on a genetic algorithm as a preferred embodiment of the three-dimensional implantation path planning system for multiple sets of brain electrodes based on a genetic algorithm.

[0048] According to the present invention, a multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithms is provided, the workflow of its modules includes:

[0049] Module M1: Collects imaging data of the brain and depth information of electrodes to generate a three-dimensional brain model;

[0050] Module M2: Based on the three-dimensional brain model, generate candidate paths; construct an objective function, analyze the candidate paths, and select the appropriate path based on the quality analysis results of the candidate paths;

[0051] Module M3: Based on real-time imagery and the objective function, optimize the corresponding path to obtain the optimal path.

[0052] This invention proposes a three-dimensional implantation path planning method for multiple sets of brain electrodes based on a genetic algorithm. This method achieves the optimal spatial distribution of multiple electrode sets through global optimization search, comprehensively considering factors such as three-dimensional vascular distribution, implantation depth, and information acquisition gain. This technology significantly improves the safety and accuracy of multi-electrode implantation, meeting the needs of high-throughput implantation.

[0053] Input data and use medical imaging data, such as MRI, CT, and PAM, to obtain the three-dimensional structure of the brain; the three-dimensional structure of the brain includes: the three-dimensional structure of brain regions and the distribution of blood vessels.

[0054] Obtain electrode implantation depth information from the user.

[0055] Path planning algorithm: based on genetic algorithm:

[0056] A. Define an objective function that takes into account constraints such as blood vessel distribution, implantation depth, and electrode spacing to optimize the electrode implantation path.

[0057] B. By simulating biological processes such as natural selection, crossover, and mutation, the optimal solution is sought among multiple alternative paths, avoiding local optima.

[0058] Online update algorithm: During the operation, the position of blood vessels and implanted electrodes is estimated by combining real-time medical images, and the electrode path is updated to cope with possible changes in blood vessels, implantation errors or other dynamic factors during the operation, so as to ensure that the path always remains optimal, thereby improving the safety and accuracy of implantation.

[0059] Medical imaging data refers to image data obtained through medical imaging technologies such as MRI, CT, and PAM. These technologies provide precise information on brain tissue structure and vascular networks. MRI (Magnetic Resonance Imaging) provides high-resolution images of the brain's soft tissue structure, used to identify brain regions and tissue boundaries; CT (Computed Tomography) provides a three-dimensional structure of the skull, helping to determine a safe implantation path; and PAM (Photoacoustic Imaging) effectively reflects the distribution of blood vessels and blood flow, providing a basis for avoiding blood vessels during implantation. By fusing these imaging data, a comprehensive three-dimensional brain model can be constructed, providing the necessary anatomical basis for electrode implantation path planning.

[0060] Genetic Algorithm: The genetic algorithm, or GA for short, is a global optimization method based on the principles of natural selection and genetics.

[0061] In this invention, a genetic algorithm is used to find the optimal path for electrode implantation in three-dimensional space.

[0062] Specifically, firstly, multiple candidate paths need to be generated, using random initialization or heuristic optimization. Secondly, the path quality is calculated based on the objective function.

[0063] The objective function is defined as follows:

[0064]

[0065] in, Let represent the union of all electrode coverage areas, and w(x) represent the information density or importance at position x. Furthermore, to ensure smooth optimization, constraints are also introduced into the objective function as soft constraints and penalty terms.

[0066]

[0067] Based on various strategies, paths with high fitness are selected. A two-point crossover strategy is adopted to improve the path optimization efficiency and introduce small-scale perturbations to avoid the algorithm getting trapped in local optima.

[0068] In other words, the two-point intersection strategy for electrode implantation refers to intersecting two implantation paths from the surface implantation site and the implantation direction to obtain a new path. The expression for the starting point of the candidate path is:

[0069] p new =(1-λ)p A +λp B

[0070] Where, p A With p B This represents two existing paths. The implantation direction is then determined using spherical linear interpolation.

[0071] Determine whether the objective function metric of the initially selected path is sorted according to the current objective function. If the result is yes, select individuals whose objective function scores are higher than the preset threshold and retain them for the next round of iteration; if the result is no, do not process them.

[0072] The so-called optimal selection strategy includes tournament selection, roulette wheel selection, and ranking selection. The fitness of different individuals in the population is measured by an objective function. A higher objective function indicates higher fitness, thereby eliminating individuals in the population that do not meet the implantation requirements.

[0073] The optimal path is output when the path optimization meets the accuracy requirements or reaches the maximum number of iterations.

[0074] The accuracy requirement is defined as the optimal fitness improvement magnitude over several consecutive generations of the population, where the improvement magnitude ΔJ < ∈ T or the maximum number of iterations is greater than T. max Then stop the calculation, assuming that the current algorithm has converged and output the optimal implantation path.

[0075] Genetic algorithms possess strong global search capabilities, enabling them to avoid local optima when faced with complex, multi-constrained path planning problems, thereby improving the accuracy and efficiency of path planning.

[0076] Objective function: The objective function is a mathematical formula used in path planning to measure the quality of the electrode path.

[0077] In this invention, the objective function is to maximize the information that the electrode can collect.

[0078] Constraints: Constraints are rules or restrictions that must be met in path planning. In this invention, the main constraints include: blood vessel avoidance, implantation depth limitation, electrode spacing, and path length.

[0079] The aforementioned vascular avoidance means that the electrode path must avoid important blood vessel areas in the brain to prevent risks such as bleeding caused by vascular damage. The vascular avoidance constraint is defined by the distance between the electrode implantation voxel and the vascular voxel set; when the distance is less than the safe distance, a certain percentage penalty is applied.

[0080]

[0081] Implantation depth limitations mean that the electrode implantation depth must be within the range set by the user to ensure that the electrode can accurately reach the target brain region. Assume the electrode implantation depth is h. i If the depth is greater than the maximum implantation depth or less than the minimum implantation depth, a certain percentage penalty will be applied.

[0082]

[0083] Electrode spacing, or the distance between electrodes, needs to meet certain requirements to avoid interference and ensure effective data acquisition. The minimum permissible spacing between electrodes is defined as s. min ,but

[0084]

[0085] Path length, or the length of the path itself, needs to be as short as possible to reduce surgical time and patient risk. The total path length for each electrode to move from the robot's zero position to the planned implantation site is defined as L. i

[0086]

[0087] Real-time medical imaging, also known as intraoperative real-time medical imaging, refers to the acquisition of real-time anatomical information of the patient's brain using specialized imaging equipment during surgery to assist surgeons in performing precise electrode implantation procedures. Common intraoperative real-time medical imaging methods include microscopic imaging and endoscopic imaging.

[0088] Microscopes provide high-magnification, clear images of the brain, enabling surgeons to clearly visualize brain surface structures during surgery. Endoscopes allow for unobstructed observation of specific surgical areas. Intraoperative registration of real-time images with preoperative 3D images allows surgeons to adjust surgical strategies in real time, ensuring safe electrode implantation.

[0089] Online learning algorithms, also known as online strategy adjustment, can be used to address dynamic or constantly changing scenarios compared to offline planning. For electrode implantation, online optimization algorithms use preoperative implantation planning as initial values ​​and already implanted electrodes as new constraints, performing incremental updates.

[0090] Assuming k electrodes have been implanted, the real-time position information obtained is p. k Then all implanted electrodes can be represented as a sequence P. k ={p1,p2,…,p k At this point, the same objective function is used for offline optimization to optimize the implantation position of the next electrode, but the penalty factor for the electrode spacing needs to be increased to avoid interference between electrodes. The results of the previous offline optimization are used as the initial values ​​for online optimization.

[0091] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0092] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A three-dimensional implantation path planning system for multiple sets of brain electrodes based on a genetic algorithm, characterized in that, include: Module M1: Collects imaging data of the brain and depth information of electrodes to generate a three-dimensional brain model; Module M2: Generates candidate paths based on the aforementioned 3D brain model; Construct an objective function, analyze the candidate paths, and select the appropriate path based on the quality analysis results of the candidate paths; Module M3: Based on real-time imagery and the objective function, optimize the corresponding path to obtain the optimal path.

2. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 1, characterized in that, In module M2, the mathematical expression of the objective function is: Where J(P) represents the objective function; P represents the position information of the electrode; ∫ represents the integral; w(x) represents the union of all electrode-covered areas; w(x) represents the information density at position x.

3. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 1, characterized in that, In module M2, the mathematical expression of the objective function is: Where J(P) represents the objective function; The union of all electrode coverage areas is represented by w(x); w(x) represents the information density at position x; P represents the position information of the electrode; ∫ represents the integral; λ1, λ2, λ3 and λ4 represent C respectively. vessel C depth C spacing With C length coefficient; C vessel Indicates vascular avoidance constraint; C depth Indicates the implantation depth limitation; C spacing Indicates electrode spacing limitation; C length Indicates path length limit.

4. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 3, characterized in that, In module M2, the candidate paths are analyzed to obtain quality analysis results, including: Module M2.1: Employs a two-point intersection strategy to select candidate paths and obtain the initial selected path; Module M2.2: Determines whether the objective function metric of the initially selected path matches the current objective function ranking. If the result is yes, individuals with objective function scores higher than the preset threshold are retained and the next iteration is performed until the preset maximum number of iterations is reached; if the result is no, they are not retained.

5. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 4, characterized in that, In module M2.1, the two-point intersection strategy involves intersecting two implantation paths from the surface implantation site and the implantation direction to obtain a new path, i.e., a candidate path; the expression for the starting point of the candidate path is: p new =(1-λ)p A +λp B Where, p new λ represents the starting point of the candidate path; λ represents the coefficient of the two-point intersection strategy; p A With p B These represent two existing implantation paths; when the two implantation paths intersect, spherical linear interpolation is performed along the implantation direction.

6. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 3, characterized in that, C vessel The mathematical expression for (P) is: Where N represents the number of electrodes to be implanted; i represents the i-th electrode; d min The minimum distance between the electrode and the blood vessel is represented by ; l represents a voxel occupied by the i-th electrode; represents the set of voxels occupied by the i-th electrode; v represents a voxel in the set of blood vessel voxels; This represents the complete set of voxels occupied by all blood vessels.

7. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 3, characterized in that, C depth The mathematical expression for (P) is: Among them, h i h represents the depth of electrode implantation. max Indicates the maximum implantation depth; h min This indicates the minimum implantation depth.

8. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 3, characterized in that, C spacing The mathematical expression for (P) is: Among them, s min Indicates the minimum permissible spacing between electrodes; p i This represents the position information of the i-th electrode; p j This represents the position information of the j-th electrode.

9. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 3, characterized in that, C length The mathematical expression for (P) is: Among them, L i L represents the total path length from the robot's zero position to the planned implantation site; max Indicates the maximum length.

10. The multi-group brain electrode three-dimensional implantation path planning system based on genetic algorithm according to claim 4, characterized in that, In module M3, the objective function metric of the initially selected path is iterated, and it is determined whether the accuracy improvement magnitude ΔJ < ∈, or the maximum number of iterations is greater than T. max If the result is yes, then stop the iteration and output the optimal path as the quality analysis result; if the result is no, then do not process it. Wherein, ∈ represents the minimum objective function lifting threshold; T max This represents the maximum number of iterations required for the genetic algorithm.

Images