Three-chamber data acquisition method, probe and acquisition system

Abstract

The method for acquiring three-compartment data includes: acquiring multiple sets of training data, wherein each set of training data includes three-compartment data and electrical data of a collection area of ​​biological tissue acquired within the same sampling period. The three-compartment data includes the extracellular space volume fraction, cell volume fraction, and blood vessel volume fraction of each voxel point in the collection area, and the electrical data is related to the electrical impedance of the biological tissue; training a computational model using multiple sets of training data, enabling the computational model to achieve the mapping between electrical data and three-compartment data; acquiring the electrical data of a target area of ​​the biological tissue, the target area being the same size and shape as the collection area; and inputting the acquired electrical data of the target area into the trained computational model to obtain the three-compartment data of the target area. This acquisition method can directly obtain three-compartment data from electrical data, reducing the complexity of three-compartment data acquisition. This invention also provides a probe and a system for acquiring three-compartment data.

Description

Technical Field

[0001] This invention relates to the medical field, and in particular to a method, probe, and acquisition system for acquiring three-compartment data. Background Technology

[0002] In the medical field, the combined measurement of electrical signals and magnetic resonance imaging (MRI) is crucial for research and diagnosis. Electrical impedance tomography (EIT), a modern imaging method, reconstructs images of internal structures by transmitting current within the body and measuring voltage distribution, and has been widely used in biomedical imaging. Magnetic resonance imaging (MRI), on the other hand, provides high-resolution three-compartment data, including the volume fractions of the extracellular space, cellular tissue, and vascular compartments. This information is essential for accurately monitoring the diffusion status of tracers and their relationship to tissue anatomy. The combination of EIT and MRI enables multimodal imaging of biological tissues. Multimodal imaging provides complementary information through different imaging modalities, allowing for a more comprehensive and accurate assessment of human structural, functional, or metabolic states, further improving the accuracy of diagnosis and research.

[0003] Currently, acquiring three-chamber data primarily relies on magnetic resonance imaging (MRI). However, MRI has many limitations, such as high cost and a cumbersome image acquisition process. Furthermore, electrical signals within biological tissues change extremely rapidly. Simultaneously acquiring electrical signals and three-chamber data while maintaining the same physiological state within the biological tissue would significantly increase the operational difficulty, not only adding to the complexity of data acquisition but also making it difficult to guarantee the stability and accuracy of the obtained data. Summary of the Invention

[0004] The purpose of this invention is to provide a method for acquiring three-compartment data, which can directly obtain three-compartment data through electrical data, thereby reducing the complexity of acquiring three-compartment data.

[0005] Another object of the present invention is to provide a probe that can introduce a tracer while acquiring electrical data of a sampling area, so as to acquire three-chamber data of the same sampling area within the same sampling period.

[0006] Another objective of this invention is to provide a three-compartment data acquisition system that can directly obtain three-compartment data through electrical data, thereby reducing the complexity of three-compartment data acquisition.

[0007] The present invention provides a method for acquiring three-compartment data, comprising: acquiring multiple sets of training data, wherein each set of training data includes three-compartment data and electrical data of a collection area of ​​biological tissue acquired within the same sampling period, wherein the three-compartment data includes the extracellular space volume fraction, cell volume fraction, and blood vessel volume fraction of each voxel point in the collection area, and the electrical data is related to the electrical impedance of the biological tissue; training a computational model using the multiple sets of training data, enabling the computational model to achieve the mapping between the electrical data and the three-compartment data; acquiring electrical data of a target area of ​​the biological tissue, wherein the target area has the same size and shape as the collection area; and inputting the acquired electrical data of the target area into the trained computational model to obtain the three-compartment data of the target area.

[0008] By training a computational model, a mapping between electrical data collected from the acquisition area and three-chamber data is achieved. Thus, by inputting the electrical data obtained from the target area into the computational model, the three-chamber data of the target area can be obtained. This method can directly obtain three-chamber data from electrical data, reducing the complexity of three-chamber data acquisition.

[0009] In another illustrative embodiment of the three-compartment data acquisition method, the method for acquiring each set of training data includes: introducing a tracer for magnetic resonance imaging into the acquisition area of ​​the biological tissue; performing magnetic resonance imaging on the acquisition area within the same sampling time period; acquiring the electrical data of the acquisition area through a probe located in the acquisition area; and obtaining the three-compartment data based on the magnetic resonance imaging results. This facilitates ensuring that each set of training data originates from the same acquisition area within the same sampling time period.

[0010] In another illustrative implementation of the three-compartment data acquisition method, the loss function of the computational model is represented by equation (1):

[0011] L(θ)=λ a L a +λ b L b +λ∥θ∥ 2 Equation (1)

[0012] in,

[0013] 'a' represents the volume fraction of the extracellular space.

[0014] b represents the cell volume fraction.

[0015] L(θ) is the loss function for the computational model, where θ represents the training parameters of the computational model.

[0016] λ a and λ b For the dynamic weights of a and b respectively,

[0017] L a and Lb These are the mean squared error losses for a and b, respectively.

[0018] λ is the regularization intensity hyperparameter.

[0019] ∥θ∥ 2 The regularization term is represented by equation (2):

[0020]

[0021] Where, θ j Let j be the training parameter of the model.

[0022] In another illustrative implementation of the three-room data acquisition method, the dynamic weight λ a and λ b Equations (3) and (4) represent the following respectively:

[0023]

[0024] in,

[0025] ∈ is a constant.

[0026] L a and L b As shown by equations (5) and (6) respectively:

[0027]

[0028] N is the number of training data sets.

[0029] a i To calculate the output value of the extracellular space volume fraction predicted by the model,

[0030] The true value of the extracellular space volume fraction is obtained from the three-compartment data of the training data, b. i To calculate the output value of the cell volume fraction predicted by the model,

[0031] The true value of cell volume fraction is obtained from the three-compartment data of the training data.

[0032] The present invention also provides a probe. The probe is configured to introduce a tracer into the acquisition area and acquire electrical data of the acquisition area when it is located in the acquisition area. This probe can introduce the tracer while acquiring electrical data of the acquisition area, thereby facilitating the acquisition of three-compartment data from the same acquisition area within the same sampling time period.

[0033] In another illustrative embodiment of the probe, the probe includes a probe body, several isolation channels, several electrodes disposed on the surface of the probe body, several wires, and a processing module. The probe body has a through-hole for introducing tracer into the acquisition area when the probe body is located in the acquisition area. Several isolation channels pass through the probe body and extend along its length. Several electrodes are arranged to form multiple layer arrays perpendicular to the length direction and are configured to acquire electrical signals through excitation measurement. Each wire passes through an isolation channel. Each electrode is connected to the processing module via a wire. The processing module can process the acquired electrical signals into electrical data and transmit the electrical data to a computational model. Integrating the electrodes into the probe body reduces the probe size and improves the stability of data acquisition.

[0034] In another illustrative embodiment of the probe, several electrodes are arranged to form three parallel layer arrays. Each layer array includes four electrodes evenly distributed circumferentially along the probe body. The electrodes in each layer array are aligned along the length of the corresponding electrodes in the adjacent layer array. This helps to improve the accuracy of data acquisition.

[0035] In another illustrative embodiment of the probe, the probe body is made of alumina ceramic, and the surface of the probe body is coated with a polyethylene glycol coating. This helps to reduce the risk of probe breakage, reduce the immune and inflammatory responses of tissue to the probe, and improve the stability of the probe.

[0036] In another illustrative embodiment of the probe, the isolation channel is made of copper, and each electrode is a circular copper sheet soldered to the surface of the probe body. This facilitates stable acquisition and transmission of electrical signals.

[0037] This invention also provides a three-compartment data acquisition system. This system includes a computing module that stores a trained computing model as described above. This three-compartment data acquisition system can directly obtain three-compartment data from electrical data, reducing the complexity of data acquisition. Attached Figure Description

[0038] The following figures are for illustrative purposes only and do not limit the scope of the invention.

[0039] Figure 1 This is a flowchart illustrating one implementation method for acquiring data from three chambers.

[0040] Figure 2 This is a flowchart illustrating one implementation method for acquiring training data.

[0041] Figure 3 This is a schematic diagram used to illustrate the sampling area.

[0042] Figure 4 and Figure 5 This is a schematic diagram illustrating the structure of the probe.

[0043] Figure 6 This is a schematic diagram used to illustrate electrical data.

[0044] Figure 7 This is a schematic diagram used to illustrate the acquired three-chamber data.

[0045] Label Explanation

[0046] 10 probes

[0047] 11 Probe Body

[0048] 111 Through Hole

[0049] 12 isolation pipes

[0050] 13 electrodes

[0051] 14 conductors

[0052] 15 processing modules

[0053] 20 Calculation Modules

[0054] S Collection Area

[0055] L is the length direction. Detailed Implementation

[0056] To provide a clearer understanding of the technical features, objectives, and effects of the invention, specific embodiments of the invention are now described with reference to the accompanying drawings, in which the same reference numerals denote the same parts.

[0057] In this document, “illustrative” means “serving as an example, illustration or description”, and any illustration or implementation described herein as “illustrative” should not be construed as a more preferred or advantageous technical solution.

[0058] Figure 1 This is a flowchart illustrating one implementation of a method for acquiring three-compartment data. See also... Figure 1 In an illustrative embodiment, the method for acquiring three-compartment data includes steps S10 to S40.

[0059] S10: Acquire multiple sets of training data. Each set of training data includes three-compartment data and electrical data of a sampling region S of biological tissue acquired within the same sampling time period. The three-compartment data includes the extracellular space volume fraction, cell volume fraction, and blood vessel volume fraction of each voxel point in the sampling region S. The sum of the three-compartment data of each voxel point is 1. The electrical data is related to the electrical impedance of the biological tissue.

[0060] Three-compartment data are acquired, for example, through magnetic resonance imaging (MRI) of biological tissues, while electrical data are acquired, for example, through the acquisition of electrical signals from biological tissues. The electrical signals of biological tissues (e.g., the brain) change rapidly; therefore, MRI and electrical signal acquisition must be performed within the same sampling period to ensure that the three-compartment data and electrical data used as training data represent biological tissues in the same physiological state. The duration of the sampling period can be determined depending on the type of biological tissue, for example, several seconds. Electrical signals are, for example, the electrical potential signals of the biological tissue, and electrical data are, for example, the potential difference between two acquisition points, which is related to the electrical impedance of the biological tissue. Those skilled in the art will recognize that electrical tomography of biological tissues can be performed using this electrical data.

[0061] Methods for obtaining training data include, for example: Figure 2 As shown in the flowchart, in Figure 2 In the illustrative implementation shown, the method for acquiring each set of training data includes S11 to S13.

[0062] S11: A tracer for magnetic resonance imaging is introduced into the acquisition area S of the biological tissue. The acquisition area S is, for example, a cubic region with an edge length of 3 cm. Figure 3 As shown by the dashed line. Tracers, for example, are... Figure 3 As shown, the tracer is introduced into the collection area S of the biological tissue through the probe 10 located in the collection area. However, it is not limited to this; in other illustrative embodiments, the tracer can also be introduced into the collection area S of the biological tissue through other means. Due to the diffusion of the tracer, the collection area S is a diffusion space centered on the tip of the probe. For ease of calculation, the collection area S can be simplified as a cubic region centered on the tip of the probe, with its centerline coinciding with the axis of the probe.

[0063] S12: Magnetic resonance imaging (MRI) is performed on the acquisition area S during the same sampling period, and electrical data of the acquisition area S is acquired through probe 10 located in the acquisition area S. In an illustrative embodiment, after the tracer is introduced into the acquisition area S, the operator performs MRI on the acquisition area S, and simultaneously acquires electrical signals through electrodes integrated on probe 10, for example, by excitation measurement, and converts the electrical signals into electrical data. The excitation measurement method uses a pair of adjacent electrodes as excitation electrodes and the remaining electrodes as measurement electrodes to acquire the electrical data of the acquisition area. In other illustrative embodiments, the excitation measurement method may also use a pair of opposing electrodes as excitation electrodes. It should be noted that a pair of opposing electrodes refers to two electrodes that are in opposite positions among all electrodes.

[0064] S13: Obtain three-chamber data based on the magnetic resonance imaging results. The three-chamber data can be obtained from the magnetic resonance imaging image according to the method disclosed in patent CN114646913B. This allows for the acquisition of multiple sets of training data, each set of training data coming from the same acquisition area during the same sampling period.

[0065] S20: Train a computational model using multiple sets of training data, enabling the computational model to map between electrical data and three-chamber data.

[0066] Multiple sets of training data are input into the computational model to obtain the mapping relationship between electrical data and three-compartment data through training. The amount of training data can be determined according to the type of biological tissue.

[0067] In the illustrative implementation, the loss function of the computational model is represented by equation (1):

[0068] L(θ)=λ a L a +λ b L b +λ∥θ∥ 2 Equation (1)

[0069] in,

[0070] L(θ) is the loss function of the computation model, where θ are the training parameters of the computation model;

[0071] λ is the regularization intensity hyperparameter;

[0072] ∥θ∥ 2 The regularization term is represented by equation (2):

[0073]

[0074] Where, θ j To calculate the j-th training parameter of the model;

[0075] a represents the volume fraction of the extracellular space;

[0076] b represents the cell volume fraction;

[0077] λ a and λ b To illustrate the dynamic weights for a and b respectively, in one illustrative implementation, the dynamic weight λ a and λ b Equations (3) and (4) represent the following respectively:

[0078]

[0079] Where ∈ is a constant;

[0080] La and L b In one illustrative implementation, L represents the mean squared error loss for a and b, respectively. a and L b As shown by equations (5) and (6) respectively:

[0081]

[0082] Where N is the number of training data sets.

[0083] a i To calculate the output value of the extracellular space volume fraction predicted by the model,

[0084] The true value of the extracellular space volume fraction was obtained from the three-compartment data of the training data.

[0085] b i To calculate the output value of the cell volume fraction predicted by the model,

[0086] The true value of cell volume fraction is obtained from the three-compartment data of the training data.

[0087] In other illustrative implementations, other loss functions can also be used for the computational model. By constraining the loss function described above, the computational model can converge quickly to complete training.

[0088] S30: Obtain electrical data of the target region of the biological tissue. The target region is the same size and shape as the collection region S. The method for obtaining the electrical data of the target region is the same as that for obtaining the electrical data of the collection region S, and will not be repeated here. It should be noted that the process of obtaining the electrical data of the target region does not require the introduction of a tracer. In addition, the relative position of the target region and the probe should also be the same as the relative position of the collection region S and the probe.

[0089] S40: Input the obtained electrical data into the trained computational model to obtain the three-chamber data of the target area.

[0090] By training a computational model, a mapping is achieved between electrical data collected from the acquisition area and three-chamber data. This allows for the acquisition of three-chamber data for the target area by inputting electrical data from the target area into the computational model. This method enables the direct acquisition of three-chamber data from electrical data, reducing the complexity of three-chamber data acquisition.

[0091] The present invention also provides a probe 10. Figure 4 and Figure 5This is a schematic diagram illustrating the probe's structure. Probe 10 is configured to introduce a tracer and acquire electrical data from the sampling area. This probe can introduce a tracer while acquiring electrical data from the sampling area, facilitating the acquisition of three-compartment data from the same sampling area within the same sampling time period.

[0092] The probe 10 includes a probe body 11, several isolation channels 12, several electrodes 13 disposed on the surface of the probe body 11, several wires 14, and a processing module 15.

[0093] The probe body 11 has a through-hole 111 for introducing tracer into the collection area S when the probe body is located in the collection area S. However, this is not a limitation; in other illustrative embodiments, the probe body can also introduce tracer into the collection area S in other ways. The probe body 11 is made of alumina ceramic, and its surface is coated with a polyethylene glycol coating. Alumina ceramic has good insulation properties and mechanical strength, making it less prone to breakage, thereby reducing the risk of probe breakage. The polyethylene glycol coating can reduce the immune and inflammatory responses of biological tissues to the probe, improving probe stability.

[0094] Several isolation pipes 12 are inserted through the probe body 11 and extend along the length direction L of the probe body 11.

[0095] Several electrodes 13 are arranged to form multiple layer arrays perpendicular to the length direction L, and can acquire electrical signals, such as potential signals, through excitation measurement. In an illustrative embodiment, the electrodes 13 are arranged in three layer arrays, each layer array including four electrodes 13 uniformly distributed circumferentially along the probe body 11. Figure 4 and Figure 5 (Only one is schematically shown in the diagram). The electrodes 13 in each layer array are aligned along the length direction L with the corresponding electrodes 13 in the adjacent layer array to avoid a decrease in accuracy at the edge regions of the acquisition area during impedance distribution reconstruction. In other illustrative embodiments, the electrodes 13 can be arranged in other numbers of layer arrays, and the number of electrodes 13 in each layer array can also be other numbers.

[0096] Each electrode 13 is a circular copper sheet welded to the surface of the probe body 11. Each electrode 13 is, for example, a copper sheet of 10 micrometers in size, to ensure high-quality signal acquisition and imaging, and good long-term stability in biological environments. Meanwhile, as... Figure 5 As shown, the circular sheet electrode 13 is integrated into the probe body 11, which can reduce the size of the probe along the length direction and reduce the size of the wound during use.

[0097] Each conductor 14 ( Figure 4 and Figure 5(Only three of them are schematically shown) are threaded through an isolation channel 12. Each electrode 13 is connected to the processing module 15 via a wire 14. The isolation channel 12 is made of copper, for example, and runs through the entire probe body 11, thereby reducing external interference and ensuring stable transmission of electrical signals.

[0098] The processing module 15 can process the acquired electrical signals into electrical data and transmit the electrical data to the computational model. In an illustrative embodiment, the processing module 15 acquires electrical signals, such as potential signals, collected by each electrode 13 through each wire 14, processes them to obtain electrical data, such as voltage values, and transmits the electrical data to the computational model, for example, through an electrical connection.

[0099] The present invention also provides a three-compartment data acquisition system. The three-compartment data acquisition system includes a computing module 20. The computing module 20 stores a trained computing model as described above. The computing module 20 is electrically connected, for example, to a processing module 15 to receive electrical data and input the electrical data into the computing model. This three-compartment data acquisition system can directly obtain three-compartment data from electrical data, reducing the complexity of three-compartment data acquisition.

[0100] The following specific embodiment illustrates how to acquire three-compartment data using a three-compartment data acquisition system. In practical use of the three-compartment data acquisition system, magnetic resonance imaging of the target area is not required, thus eliminating the need for introducing tracers into the target area. Only the electrical data of the target area needs to be acquired using a probe and input into the computational model stored in the computation module 20 for mapping. Therefore, in practical use, the probe can be, for example... Figure 4 and Figure 5 The probe 10 shown can also be other probes that do not have through holes 111.

[0101] The biological tissue is the brainstem region. A probe 10 is inserted into the target area of ​​the brainstem region. The probe 10 integrates 12 electrodes 13, arranged in three layer arrays, for example, named layer arrays A, B, and C. The electrodes 13 in each layer array are named, for example, electrodes A1 to A4, B1 to B4, and C1 to C4. An excitation measurement method is used. First, for example, two adjacent electrodes 13, A1 and A2, are used as excitation electrodes, and the remaining 10 electrodes 13 are used as measurement electrodes. During the excitation measurement process, an excitation current is passed through electrodes A1 and A2. The processing module 15 collects 10 potential signals output from the remaining 10 electrodes, processes them, and obtains 9 voltage values, i.e., 9 electrical data. This process is repeated, sequentially switching A2 and A3, A3 and A4, A4 and A1, B1 and B2… as excitation electrodes to complete a complete excitation measurement process. This process yields 108 electrical data points, i.e., electrical data, such as… Figure 6 As shown. Figure 6The horizontal axis represents the serial number of the electrical data, and the vertical axis represents the voltage, with the unit being mV.

[0102] The probe transmits the electrical data acquired from the brainstem region to the computing module 20. The computing module 20 receives the electrical data and inputs it into the trained computing model. The computing model then calculates the following... Figure 7 The data for the three chambers is shown. Figure 7 In the image, from left to right, the volume fraction distributions of the extracellular space, cells, and blood vessels in the brainstem region are shown. At each voxel point, the sum of the volume fractions of the extracellular space, cells, and blood vessels is 1.

[0103] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0104] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention and are not intended to limit the scope of protection of the present invention. All equivalent implementation schemes or modifications made without departing from the spirit of the present invention, such as combinations, divisions or repetitions of features, should be included within the scope of protection of the present invention.

Claims

1. A method for acquiring three-room data, characterized in that, include: Multiple sets of training data are acquired, wherein each set of training data includes three-compartment data and electrical data of a collection area of ​​biological tissue acquired within the same sampling period. The three-compartment data includes the extracellular space volume fraction, cell volume fraction and blood vessel volume fraction of each voxel point in the collection area. The electrical data is related to the electrical impedance of the biological tissue. A computational model is trained using multiple sets of training data, enabling the computational model to achieve the mapping between the electrical data and the three-chamber data; Acquire electrical data of a target region of a biological tissue, wherein the target region is the same size and shape as the acquisition region; and The obtained electrical data of the target area is input into the trained computational model to obtain the three-chamber data of the target area; wherein The loss function of the computational model is expressed by equation (1): ） in, 'a' represents the volume fraction of the extracellular space. b represents the cell volume fraction. The loss function of the computational model, where These are the training parameters for the computational model. and For the dynamic weights of a and b respectively, These are the mean squared error losses for a and b, respectively. This is the hyperparameter for regularization strength. The regularization term is represented by equation (2): in, For the computational model of the first One training parameter; Dynamic weights and Equations (3) and (4) represent the following respectively: in, It is a constant. As shown by equations (5) and (6) respectively: in, N is the number of training data sets. The output value is the extracellular space volume fraction predicted by the computational model. The true value of the extracellular space volume fraction is obtained from the three-compartment data of the training data. The output value of the cell volume fraction predicted by the computational model. The true value of the cell volume fraction is obtained from the three-compartment data of the training data.

2. The method for acquiring three-chamber data as described in claim 1, characterized in that, The methods for obtaining the training data in each group include: A tracer for magnetic resonance imaging is introduced into the acquisition area of ​​a biological tissue. Magnetic resonance imaging was performed on the acquisition area during the same sampling period, and electrical data of the acquisition area were acquired through a probe located in the acquisition area. The data for the three chambers were obtained based on the results of magnetic resonance imaging.

3. A probe, characterized in that, Using the three-compartment data acquisition method as described in claim 1, the probe is configured to introduce a tracer into the acquisition region (S) and acquire electrical data of the acquisition region (S) when it is located in the acquisition region (S).

4. The probe as described in claim 3, characterized in that, The probe includes: A probe body (11) has a through hole (111) for introducing tracer into the acquisition area (S) when the probe body (11) is located in the acquisition area. Several isolation channels (12) are embedded within the probe body (11). Several electrodes (13) are disposed on the surface of the probe body (11). The electrodes (13) are arranged to form multiple layer arrays perpendicular to the length direction (L) of the probe body (11) and are configured to acquire electrical signals by means of excitation measurement. Several wires (14), each wire (14) passing through one of the isolation conduits (12) and electrically connected to one of the electrodes (13), and A processing module (15) is provided, wherein each of the electrodes (13) is connected to the processing module (15) via a wire (14). The processing module (15) is capable of processing the acquired electrical signals into electrical data and transmitting the electrical data to the computational model.

5. The probe as described in claim 4, characterized in that, Several electrodes (13) are arranged to form three mutually parallel layer arrays, each layer array including four electrodes (13) evenly distributed circumferentially along the probe body (11), and the electrodes (13) in the three layer arrays are aligned along the length direction (L).

6. The probe as described in claim 4, characterized in that, The probe body (11) is made of alumina ceramic, and the surface of the probe body (11) is covered with a polyethylene glycol coating.

7. The probe as described in claim 4, characterized in that, The isolation conduit (12) is made of copper, and each of the electrodes (13) is a circular copper sheet welded to the surface of the probe body (11).

8. A three-room data acquisition system, comprising a computing module (20), wherein the computing module (20) stores a trained computing model as described in any one of claims 1 to 2.

Images