Neuron feeling model-based tactile map construction method and system

Abstract

The invention discloses a touch map construction method and system based on a neuron feeling model. The construction method comprises the following steps: acquiring a structure image and a function image; the structure image and the function image are preprocessed, a touch brain area is positioned, and the touch brain area comprises a voxel space and a vertex space; constructing a tactile stimulation frame model; constructing a neuron feeling model in a voxel space and a vertex space based on the tactile stimulation frame model; decoding the neuron feeling model to obtain tactile stimulation information; and constructing a tactile map based on the tactile stimulation information. According to the method, the problem of quantitative coupling of the tactile stimulation information and the receptive field of the tactile neurons can be solved, and finer tactile map construction is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of image data processing, and particularly relates to a method and system for constructing a tactile map based on a neuron receptive model. Background Art

[0002] As a basic sensory modality, touch is a key way for humans to perceive the world and interact with the environment. It enables us to perceive the texture, shape, hardness, and surface details of objects, which is crucial for our daily life and work. In the field of neuroscience, the neural mechanism of touch has always been the focus of attention because it reveals how the brain processes tactile information from the outside world and converts it into meaningful perceptual experiences. In the early stage of touch research, scientists mainly constructed tactile maps through classical neurophysiological experiments and anatomical studies. These research methods usually involved precisely stimulating the nervous system of experimental animals while recording the responses of neurons, or inferring the processing path of tactile information through direct observation of brain tissue. Through these methods, preliminary tactile perception areas were mapped. However, these traditional research methods have obvious limitations. First, they often can only provide general structural information about the tactile system, that is, which brain regions are associated with tactile perception. Although this information is important, it is static and cannot reveal the dynamic process of tactile information processing. Second, due to technical limitations, these methods are difficult to penetrate into the cellular and molecular levels, so they cannot detail the microscopic mechanism of tactile perception. For example, they cannot accurately describe the interactions between neurons, the molecular basis of signal transmission, and the encoding and decoding processes of tactile information in the brain.

[0003] With the development of technology, functional magnetic resonance imaging (fMRI) has been widely used in the construction of tactile maps. The human brain tactile map aims to reveal the neural activity patterns and brain region function distributions related to tactile perception through means such as neuroimaging technology and electrophysiological technology. Through specific tactile stimulation experiments, researchers can observe the activities of brain regions related to touch in the brain, thereby gaining a deeper understanding of the neural basis of tactile perception. However, these methods still have certain limitations. For example, the spatial resolution of fMRI is limited, making it difficult to accurately locate the activities of individual neurons or small groups of neurons. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a method and system for constructing a tactile map based on a neuron receptive model, which can more accurately locate the activities of individual neurons or small groups of neurons.

[0005] The present invention provides a method for constructing a tactile map based on a neuron receptive model, including:

[0006] Obtaining a structural image and a functional image;

[0007] Preprocess the structural image and the functional image to locate the somatosensory brain region, where the somatosensory brain region includes a voxel space and a vertex space;

[0008] Construct a somatosensory stimulation frame model;

[0009] In the voxel space and the vertex space, construct a neuron receptive model based on the somatosensory stimulation frame model;

[0010] Decode the neuron receptive model to obtain somatosensory stimulation information;

[0011] Construct a somatosensory map based on the somatosensory stimulation information.

[0012] Optionally, obtaining the structural image and the functional image includes:

[0013] Set a somatosensory stimulation task based on visual attention;

[0014] Perform magnetic resonance scanning based on the somatosensory stimulation task to obtain the scanned structural image and functional image.

[0015] Optionally, preprocessing the structural image includes:

[0016] Perform spatial normalization on the structural image;

[0017] Perform gray matter segmentation on the normalized structural image to obtain a segmentation result;

[0018] Perform gray matter dilation on the segmentation result to obtain a processed structural image.

[0019] Optionally, constructing a neuron receptive model in the voxel space and the vertex space based on the somatosensory stimulation frame model includes:

[0020] In the voxel space, construct a voxel-level neuron receptive model based on the somatosensory stimulation frame model;

[0021] In the vertex space, construct a vertex-level neuron receptive model based on the somatosensory stimulation frame model.

[0022] Optionally, constructing a voxel-level neuron receptive model in the voxel space based on the somatosensory stimulation frame model includes:

[0023] Construct a voxel neuron receptive model;

[0024] Perform linear analysis on the voxel neuron receptive model, and fit the analyzed voxel neuron receptive model to obtain a voxel-level neuron receptive model.

[0025] Optionally, in the vertex space, constructing a neuron receptive model at the vertex level based on the tactile stimulation frame model includes:

[0026] Construct a vertex neuron receptive model;

[0027] Perform linear analysis on the vertex neuron receptive model, and fit the analyzed vertex neuron receptive model to obtain a neuron receptive model at the vertex level.

[0028] The present invention also provides a tactile map construction system based on a neuron receptive model, including: a tactile stimulation induction module, a data acquisition and processing module, and a tactile reconstruction and display module;

[0029] The tactile stimulation induction module is used to design a tactile stimulation task;

[0030] The data acquisition and processing module is used to acquire and process the front-end functional imaging and structural imaging data based on the tactile stimulation task;

[0031] The tactile reconstruction module is used to decode the neuron receptive model based on the processed data and construct a tactile map.

[0032] Compared with the prior art, the present invention has the following advantages and technical effects:

[0033] The present invention can solve the problem of quantitative coupling between tactile stimulation information and the receptive field of tactile neurons, and realizes a more refined tactile map construction, including tactile maps in two dimensions between fingers and within fingers, and can describe the tactile activation maps of each finger joint. Description of the Drawings

[0034] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0035] Figure 1 is a flowchart of a method for constructing a tactile map based on a neuron receptive model according to an embodiment of the present invention;

[0036] Figure 2 is a schematic diagram of an experimental paradigm for inducing tactile stimulation based on visual attention according to an embodiment of the present invention;

[0037] Figure 3 is a schematic diagram of a binary tactile stimulation frame model of a square image designed based on the experimental paradigm according to an embodiment of the present invention;

[0038] Figure 4 is a schematic diagram of a tactile map according to an embodiment of the present invention;

[0039] Figure 5Schematic diagram of the tactile map construction system based on the neuron receptive model according to an embodiment of the present invention. Detailed implementation manners

[0040] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0041] It should be noted that the steps shown in the flowchart of the drawings may be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than here.

[0042] This embodiment proposes a method for constructing a tactile map based on a neuron receptive model, as Figure 1 shown, specifically including the following steps:

[0043] Obtain a structural image and a functional image;

[0044] Preprocess the structural image and the functional image, and locate the tactile brain region, where the tactile brain region includes a voxel space and a vertex space;

[0045] Construct a tactile stimulus frame model;

[0046] In the voxel space and the vertex space, based on the tactile stimulus frame model, construct a neuron receptive model;

[0047] Decode the neuron receptive model to obtain tactile stimulus information;

[0048] Based on the tactile stimulus information, construct a tactile map.

[0049] Specifically, this embodiment includes:

[0050] S1: Based on visual attention, design a tactile stimulus-induced experiment, and scan the functional imaging and structural images of the subject;

[0051] S2: Preprocess the original brain images, and further design a binary tactile stimulus frame of a square image;

[0052] S3: In the voxel space, construct a voxel-level neuron receptive model;

[0053] S4: On the cortical surface space, construct a vertex-level neuron receptive model;

[0054] S5: Decode the neuron receptive model, and further perform tactile stimulus reconstruction.

[0055] Furthermore, obtaining a structural image and a functional image includes:

[0056] Set a tactile stimulation task based on visual attention;

[0057] Perform magnetic resonance scanning based on the tactile stimulation task to obtain the structural image and functional image after scanning.

[0058] Specifically, in step S1, this embodiment designed a tactile stimulation-induced experiment based on visual attention. The stimulation type was mechanical vibration stimulation, and the stimulation frequency was 21 Hz. During the experiment, the hand was subdivided into 23 tactile stimulation areas, including the proximal (finger root, Dx3), middle (interphalangeal, Dx2), and distal (fingertip, Dx1) phalanx areas of the thumb (D1), index finger (D2), middle finger (D3), ring finger (D4), and little finger (D5) of the left hand, as well as nine stimulation areas on the palm (P), as Figure 2 shown. The specific experimental stimulation process was as follows: Stimulate each finger in turn from the thumb to the ring finger (from the distal phalanx to the proximal phalanx in turn), and then stimulate the palm. A total of 23 positions were stimulated in the whole process, and each position was stimulated for 8 seconds (continuously stimulated for 7 seconds, with an interval of 1 second to avoid adaptation). Then move to the next position. After completing the stimulation of 14 positions on the fingers, an 8-second rest period was set, followed by the stimulation of 9 positions on the palm, and then a 16-second rest. Thus, a complete stimulation cycle was formed, and the whole cycle took 216 seconds. The task-state fMRI scan consisted of 8 such stimulation cycles, with a total duration of 1672 seconds. To ensure that the attention state of the participants was controlled, a "+" visual stimulation was introduced during the experiment, and the subjects kept their eyes open and stared at the cross in the center of the screen throughout the task-state scanning stage. Further, during the tactile stimulation task, magnetic resonance scanning included 7T task-state functional magnetic resonance imaging and high-resolution structural images of T1-weighted imaging.

[0059] Furthermore, the preprocessing of the structural image includes:

[0060] Perform spatial normalization on the structural image;

[0061] Perform gray matter segmentation on the normalized structural image to obtain the segmentation result;

[0062] Perform gray matter inflation on the segmentation result to obtain the processed structural image.

[0063] Furthermore, the preprocessing of the functional image includes:

[0064] Perform time correction, spatial correction, filtering, and smoothing on the functional image to obtain the preprocessed functional image.

[0065] Specifically, in step S2, preprocessing of the structural data is performed, including: To ensure the unity and accuracy of the analysis, the individual brain structural images of each subject need to be mapped to the standard space to eliminate the differences between subjects and facilitate cortical activation analysis. Given that the intensity inhomogeneity of magnetic resonance images will greatly reduce the accuracy of segmentation and registration. Therefore, intensity inhomogeneity correction is carried out, mainly including image cleaning, brain structure extraction, white matter detection, and bias field correction within white matter voxels. To accurately estimate the boundaries between white matter and gray matter and between gray matter and cerebrospinal fluid, advanced segmentation techniques are used. Through this series of processes, high-quality segmentation results are obtained, which are used to create high-quality cortical reconstructions, and at the same time, gray matter inflation is performed to visualize high-resolution functional data. The data processing process of task-based functional magnetic resonance imaging includes: slice scan time correction, head motion correction, and high-pass filtering in the time domain, and spatial smoothing. Then, the processed functional images are registered with the structural images. Further, a binary tactile stimulation frame of a square image is designed. During the experiment of this study, a total of 23 tactile stimulation points are designed. Therefore, these positions are defined in the two-dimensional hand space, and the coordinate scale of each dimension is limited within ±11.5 units to ensure that the 23 stimulation points on the hand can be covered. Different from the commonly used polar coordinate system in the visual space, the hand space defined in this study uses the Cartesian coordinate system, where the x-axis (representing the dimension between fingers) is divided into 23 intervals corresponding to different hand stimulation positions, and the y-axis is not refined in this model construction. Therefore, a binary tactile stimulation frame model of a square image as shown in Figure 3 is constructed. This representation in the binary stimulation frame may be simpler than the actual stimulation used during the experiment because it abstracts from the finer details within the stimulated area.

[0066] Furthermore, in the voxel space and vertex space, based on the tactile stimulation frame model, constructing a neuronal receptive model includes:

[0067] In the voxel space, based on the tactile stimulation frame model, constructing a voxel-level neuronal receptive model;

[0068] In the vertex space, based on the tactile stimulation frame model, constructing a vertex-level neuronal receptive model.

[0069] Furthermore, in the voxel space, based on the tactile stimulation frame model, constructing a voxel-level neuronal receptive model includes:

[0070] Constructing a voxel neuronal receptive model;

[0071] Performing a linear analysis on the voxel neuronal receptive model, fitting the analyzed voxel neuronal receptive model, and obtaining a voxel-level neuronal receptive model.

[0072] Specifically, in step S3, in the voxel space, for each voxel included in the analysis, a separate linear analysis is performed for the neuronal receptive model of each voxel, and the time course of the model is used as the main predictor (predictor of interest) of the linear analysis design matrix X, thereby allowing different beta weights to be fitted for the baseline level. The time series data of the voxel is separately converted into percentage signal change values in each run and then concatenated to provide the observed (to be explained) data y for the linear analysis. A linear analysis is performed on each neuronal receptive model, and the explained variance R 2 value is recorded. The linear analysis model with the highest explained variance value is assigned as the neuronal receptive model of the voxel under consideration. This process is repeated for each voxel included in the analysis. The parameters of the best-fitting model are stored in the tactile activation map that provides the final output of the fitting process, that is, for each analyzed voxel, the position (x, y values) of the neuronal receptive model in the visual field and the receptive field size of the model are recorded in the tactile activation map. The neuronal receptive fitting models at the voxel level are integrated to form a population of neuronal receptive models.

[0073] Furthermore, in the vertex space, based on the tactile stimulation frame model, constructing a neuronal receptive model at the vertex level includes:

[0074] Constructing a vertex neuronal receptive model;

[0075] Performing a linear analysis on the vertex neuronal receptive model, fitting the analyzed vertex neuronal receptive model, and obtaining a neuronal receptive model at the vertex level.

[0076] Specifically, in step S4, in the cortical surface space, for each vertex included in the analysis, a separate linear analysis is performed for the neuronal receptive model of each vertex, and the time course of the model is used as the main predictor (predictor of interest) of the linear analysis design matrix X, thereby allowing different beta weights to be fitted for the baseline level. The time series data of the vertex is separately converted into percentage signal change values in each run and then concatenated to provide the observed (to be explained) data y for the linear analysis. A linear analysis is performed on each neuronal receptive model, and the explained variance R 2 quantity is recorded. The linear analysis model with the highest explained variance value is assigned as the neuronal receptive model of the vertex under consideration. This process is repeated for each vertex included in the analysis. The parameters of the best-fitting model are stored in the tactile map of the cortical surface that provides the final output of the fitting process, that is, for each analyzed vertex, the position (x, y values) of the neuronal receptive model in the visual field and the receptive field size are recorded in the tactile map of the cortical surface. The neuronal receptive fitting models at the vertex level are integrated to form a population of neuronal receptive models.

[0077] In step S5, in the voxel and cortical surface space, after obtaining the population neuron receptive model, the neuron receptive model visualizes the receptive field positions and sizes of voxels (and vertices) in the tactile input space to achieve the decoding of tactile stimuli. The decoding method is manifested as visualizing the model parameters of specific voxels or vertices at corresponding positions in the tactile brain region space, or visualizing the index of "intensity" weighted by the voxel activation level in the tactile brain region space according to the activation intensity of the voxels. When integrating this information from the tactile cortex, the tactile stimuli of the participant can be reconstructed. For the stimulus reconstruction method, by projecting the neuron receptive model of the activated voxels or vertices into the tactile receptive field, the activation distribution pattern is converted into an image of the tactile receptive field; the projection intensity ("brightness") of the neuron receptive model at specific positions in the tactile receptive field depends on the activity amplitude of the voxels (or vertices). This process is actually "decoding" the activity pattern in the visual cortex in the visual field representation and reconstructing the tactile image of the stimulus. Threshold limiting is performed on the intensity values of the reconstructed image. The intensity values are scaled from 0.0 to 1.0 (0.0 - 1.0) according to the minimum and maximum reconstructed values in the reconstructed image. In this embodiment, the intensity value is set to 0.5, and the intensity range is mapped to the standard blue, green, yellow, and red ranges. If the image is thresholded, the result can also be binarized, that is, pixels with values higher than the threshold will be displayed in one color (black), while pixels lower than the threshold will be displayed in another color (white background), and finally a tactile map is constructed, as Figure 4 shown.

[0078] As Figure 5 shown, this embodiment also provides a tactile map construction system based on the neuron receptive model, including: a tactile stimulus induction module, a data acquisition and processing module, and a tactile reconstruction display module;

[0079] The tactile stimulus induction module is used to design tactile stimulus tasks;

[0080] The data acquisition and processing module is used to collect and process the front-end functional imaging and structural imaging data based on the tactile stimulus tasks;

[0081] The tactile reconstruction module is used to decode the neuron receptive model based on the processed data and construct a tactile map.

[0082] Specifically, this embodiment also provides a tactile map construction system based on the neuron receptive model, including a tactile stimulus induction module for designing more refined tactile stimulus paradigms and tactile stimulus means; a data acquisition and processing module for collecting and processing the front-end functional imaging and structural imaging data; a tactile reconstruction display module for displaying the tactile stimulus positions and stimulus sizes after decoding the neuron receptive model, as well as displaying the tactile map.

[0083] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for constructing a tactile map based on a neuron perception model, characterized in that Including: Obtaining a structural image and a functional image; Preprocessing the structural image and the functional image to locate the somatosensory brain region, where the somatosensory brain region includes a voxel space and a vertex space; Constructing a somatosensory stimulation frame model; In the voxel space and the vertex space, constructing a neuronal receptive model based on the somatosensory stimulation frame model; Decoding the neuronal receptive model to obtain somatosensory stimulation information; Constructing a somatosensory map based on the somatosensory stimulation information.

2. The method for constructing a tactile map based on a neuron perception model according to claim 1, wherein Obtaining a structural image and a functional image includes: Setting a somatosensory stimulation task based on visual attention; Performing magnetic resonance scanning based on the somatosensory stimulation task to obtain the scanned structural image and functional image.

3. The method for constructing a tactile map based on a neuron perception model according to claim 1, wherein Preprocessing the structural image includes: Performing spatial normalization on the structural image; Performing gray matter segmentation on the normalized structural image to obtain a segmentation result; Performing gray matter dilation on the segmentation result to obtain the preprocessed structural image.

4. The method for constructing a tactile map based on a neuron receptive model according to claim 3, wherein Preprocessing the functional image includes: Performing temporal correction, spatial correction, filtering, and smoothing on the functional image to obtain the preprocessed functional image.

5. The method for constructing a tactile map based on a neuron perception model according to claim 1, wherein In the voxel space and the vertex space, constructing a neuronal receptive model based on the somatosensory stimulation frame model includes: In the voxel space, constructing a voxel-level neuronal receptive model based on the somatosensory stimulation frame model; In the vertex space, constructing a vertex-level neuronal receptive model based on the somatosensory stimulation frame model.

6. The method for constructing a tactile map based on a neuron receptive model according to claim 5, characterized in that, In the voxel space, constructing a voxel-level neuronal receptive model based on the somatosensory stimulation frame model includes: Constructing a voxel neuronal receptive model; Perform a separate linear analysis for the neuronal receptive model of each voxel and record the explained variance R 2 value. The linear analysis model with the highest explained variance value is assigned as the neuronal receptive model for the voxel under consideration, and a voxel-level neuronal receptive model is obtained.

7. The method for constructing a tactile map based on a neuron perception model according to claim 5, wherein In the vertex space, constructing a vertex-level neuronal receptive model based on the somatosensory stimulation frame model includes: Constructing a vertex neuronal receptive model; Perform a separate linear analysis for the neuronal receptive model of each vertex and record the explained variance R 2 value. The linear analysis model with the highest explained variance value is assigned as the neuronal receptive model for the vertex under consideration, obtaining the neuronal receptive model at the vertex level.

8. The tactile map construction system based on the neuron perception model is characterized in that, Including: A somatosensory stimulation induction module, a data acquisition and processing module, and a somatosensory reconstruction display module; The somatosensory stimulation induction module is used to design a somatosensory stimulation task; The data acquisition and processing module is used to acquire and process the front-end functional imaging and structural imaging data based on the somatosensory stimulation task; The somatosensory reconstruction module is used to decode the neuronal receptive model based on the processed data and construct a somatosensory map.

Images