Multi-modal joint evaluation method for repairing effect of eye-brain assembly

Abstract

The invention discloses a multi-modal joint evaluation method for the repairing effect of an eye-brain assembly, and belongs to the technical field of biomedical engineering and neuroscience. The method aims at solving the problems that an existing evaluation means cannot reflect function integration in real time in a living body state, a specific index is lacked to distinguish effective remodeling and disordered growth, and a single behavioral test is prone to interference, so that false positive is caused. Specific visual stimulation is presented to a transplantation receptor, local field potential of a primary visual cortex is synchronously collected, and Low Gamma frequency band signals are extracted to calculate direction selectivity indexes and brain network topological attributes; meanwhile, a hierarchical visual behavioral test for sensing spatial cognition from basic light is carried out; and finally, constructing a binary decision matrix based on thresholding scores of electrophysiology and behavioral indexes, and outputting a four-classification quantitative evaluation conclusion including effective functional repair, thereby realizing accurate and standardized judgment of the visual cortex repair effect.

Description

Technical Field

[0001] This invention belongs to the fields of biomedical engineering and neuroscience, and in particular relates to a multimodal joint evaluation method for the repair effect of the eye-brain assembly. Background Technology

[0002] In the fields of biomedical engineering, neuroscience, and brain-computer interface evaluation, assessing the effectiveness of eye-brain assemblies (such as organoids or stem cell grafts) in repairing visual cortex injuries has always been a key technical challenge. Currently, three main technical approaches are relied upon, but each has significant limitations in specific applications. First, morphological evaluation methods, such as immunohistochemistry or neural tracing, rely on the observation of sections of ex vivo tissue. This method can only statically assess the survival status of the graft and the structural integrity of axonal projections, but cannot characterize, in real-time and dynamically, whether the repaired neurons possess true electrical signal transmission and information processing capabilities in live animal models. In practical applications, this makes it difficult for researchers to distinguish whether the graft has formed an effective functional neural circuit or has merely undergone nonfunctional, disordered growth, thus making it impossible to accurately determine the substantial effect of the repair. Second, conventional electrophysiological evaluation indicators, such as changes in the amplitude and latency of visual evoked potentials, or broadband EEG energy analysis, face challenges in practical applications due to their lack of sufficient specificity. Damaged brain regions are often accompanied by abnormal or disordered spontaneous neural activity, and nonspecific light reflexes may also induce similar signals. These interferences make it difficult for existing electrophysiological indicators to accurately quantify the ability of repaired neural tissue to specifically extract visual features (such as direction and spatial frequency), easily leading to false positives in assessment results and misjudging meaningless electrical activity as functional recovery. Furthermore, relying solely on visual behavioral tests, such as the light-dark box experiment, results in significant errors and cannot distinguish the neural circuit levels of functional recovery. Such tests are highly susceptible to interference from the experimental animal's own motor impairments, fluctuations in anxiety levels, or preserved subcortical reflex pathways (such as the superior colliculus pathway). In actual testing, the animal may only exhibit basic light perception abilities, but this does not prove that its higher visual cortical functions (such as depth perception, pattern recognition, or spatial navigation) have been restored, leading to ambiguous assessment conclusions and an inability to construct a complete chain of evidence from the electrophysiological activity of microscopic neural circuits to the macroscopic visual perception behavior of the animal. The individual limitations of these existing technologies, as well as the lack of an effective joint verification mechanism among them, make it extremely difficult to conduct accurate, quantitative, and standardized assessments of the functional repair effects after eye-brain assembly transplantation. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention proposes a multimodal joint evaluation method for the repair effect of the eye-brain assembly, thereby resolving the issues existing in the prior art.

[0004] In a first aspect, to achieve the above objectives, the present invention provides a multimodal joint evaluation method for the repair effect of the eye-brain assembly, comprising the following steps: Visual stimuli with different orientations were presented to experimental subjects who had received eye-brain assemblies, and local field potential signals of their primary visual cortex were collected simultaneously. The local field potential signal is processed to extract neural oscillation signals of a specific frequency band, and based on the oscillation signals, electrophysiological indicators for characterizing the selectivity of visual features and topological indicators for characterizing the degree of brain network integration are calculated. The experimental subjects underwent hierarchical visual behavioral tests, from simple to complex, to obtain behavioral indicators corresponding to different levels of visual perception. Based on preset thresholds, the electrophysiological and behavioral indicators are binarized and scored respectively. A joint decision matrix is ​​constructed based on the scoring results, and a four-category evaluation conclusion on the repair effect of the eye-brain assembly is output.

[0005] Optionally, the visual stimulus is a drifting sinusoidal grating presented on the screen, and the presentation paradigm of the visual stimulus is an event-related design. Each trial includes a visual stimulus presentation period and a gray screen interval period of equal brightness. Stimuli from different orientations are presented in a pseudo-random order for multiple cycles.

[0006] Optionally, the specific frequency band is the Low Gamma band; the extraction process includes: bandpass filtering the local field potential signal to obtain the neural oscillation signal of the Low Gamma band, and calculating the relative power change of the oscillation signal relative to the baseline during the presentation of the visual stimulus.

[0007] Optionally, the electrophysiological index used to characterize the selectivity of visual features is the direction selectivity index; the calculation process includes: calculating the direction selectivity index by vector synthesis based on the relative power values ​​of the response of the Low Gamma band neural oscillation signal under different visual stimulus orientations.

[0008] Optionally, topological metrics used to characterize the degree of brain network integration include global efficiency and clustering coefficient; the calculation process includes: constructing a functional connectivity network based on low-gamma band neural oscillation signals acquired through multiple channels using a synchronous likelihood estimation algorithm, and then calculating the global efficiency and clustering coefficient of the functional connectivity network.

[0009] Optional, hierarchical visual behavioral tests include: the light and dark box test for testing basic light perception, the visual cliff test for testing depth perception, and the visual water maze test for testing graphic spatial discrimination.

[0010] Optionally, the process of binarizing the electrophysiological indicators includes: setting judgment thresholds for the directional selectivity index, the global efficiency, and the clustering coefficient, counting the number of indicators that reach the corresponding thresholds, and determining that the electrophysiological dimension score meets the standard when the number of indicators that reach the thresholds is not less than the set value.

[0011] Optionally, the joint decision matrix is ​​a two-dimensional matrix, where one dimension represents whether the electrophysiological index score meets the standard, and the other dimension represents whether the behavioral index score meets the standard; the four-category assessment conclusion includes: both electrophysiological and behavioral scores meet the standard, corresponding to effective functional repair; only the electrophysiological score meets the standard, corresponding to invalid connectivity; only the behavioral score meets the standard, corresponding to non-cortical-dependent compensation; and neither meets the standard, corresponding to no repair.

[0012] Secondly, the present invention also provides a computer terminal device, comprising: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the multimodal joint evaluation method for the repair effect of the eye-brain assembly in the first aspect described above.

[0013] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the multimodal joint evaluation method for the repair effect of the eye-brain assembly described in the first aspect are implemented.

[0014] Compared with the prior art, the present invention has the following advantages and technical effects: This invention provides a multimodal joint evaluation method for the repair effect of the eye-brain assembly. By integrating multimodal data from microscopic electrophysiology and macroscopic behavioral science, it achieves a multidimensional quantitative characterization of the repair effect of the visual cortex. This invention combines high spatiotemporal resolution extracellular electrophysiological signals with hierarchical visual behavioral indicators to construct a complete evaluation link from microscopic neural circuits to macroscopic visual perception, overcoming the limitations of single evaluation methods. By establishing a core evaluation index based on the Low Gamma band directional selectivity index, it can specifically identify whether neurons have formed an ordered functional network that finely processes visual features, significantly improving the specificity and sensitivity of the evaluation. Simultaneously, functional brain network analysis is introduced to quantitatively reveal the degree of integration between the graft and the host visual network. Finally, by constructing a binary joint decision matrix of microscopic and macroscopic data for cross-validation and logical judgment, interference factors such as non-cortical-dependent compensation and invalid connections are effectively eliminated. This establishes a standardized, quantitative, and definitive method for evaluating the functional repair of the eye-brain assembly, significantly improving the accuracy, robustness, and universality of the evaluation. Attached Figure Description

[0015] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a visual stimulus paradigm diagram of an embodiment of the present invention; Figure 2 This is a flowchart illustrating a multimodal joint evaluation method for the repair effect of an eye-brain assembly according to an embodiment of the present invention. Figure 3 This is a flowchart of signal processing and feature extraction based on the Low Gamma band in an embodiment of the present invention. Detailed Implementation

[0016] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0017] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0018] like Figure 2 As shown, this embodiment provides a multimodal joint evaluation method for the repair effect of the eye-brain assembly, including: Visual stimuli with different orientations were presented to experimental subjects who had received eye-brain assemblies, and local field potential signals of their primary visual cortex were collected simultaneously. The local field potential signal is processed to extract neural oscillation signals of a specific frequency band, and based on the oscillation signals, electrophysiological indicators for characterizing the selectivity of visual features and topological indicators for characterizing the degree of brain network integration are calculated. The experimental subjects underwent hierarchical visual behavioral tests, from simple to complex, to obtain behavioral indicators corresponding to different levels of visual perception. Based on preset thresholds, the electrophysiological and behavioral indicators are binarized and scored respectively. A joint decision matrix is ​​constructed based on the scoring results, and a four-category evaluation conclusion on the repair effect of the eye-brain assembly is output.

[0019] Furthermore, the visual stimulus is a drifting sinusoidal grating presented on the screen, and the presentation paradigm of the visual stimulus is an event-related design. Each trial includes a visual stimulus presentation period and a gray screen interval period of equal brightness. The stimuli from different directions are presented in a pseudo-random order for multiple cycles.

[0020] Specifically, the implementation process of this embodiment includes: Step S1: Microelectrode array implantation and signal acquisition under specific visual stimulation paradigms This step aims to establish a physical-level "visual input-neural output" interface, and the specific implementation process is as follows: ① First, the microelectrode array was positioned and implanted. The recipient mice were anesthetized and fixed on a stereotaxic apparatus to expose the skull surface. Based on the pre-defined coordinates for the eye-brain assembly transplantation surgery, the center of the primary visual cortex (V1) transplantation area was determined, specifically 4.0 mm posterior to the anterior fontanelle (AP: -4.0 mm) and 2.7 mm lateral to the median suture (ML: 2.7 mm). A craniotomy was performed at this location. After removing the dura mater, a multi-channel silicon probe microelectrode array (16-channel linear probe) was vertically and slowly inserted into the brain tissue using a micromanipulator until the electrode contacts completely covered the graft depth, ensuring the acquisition of local field potential (LFP) signals within the graft.

[0021] ② Secondly, a visual stimulus presentation system was constructed. The mouse was placed 20cm directly in front of the monitor screen (37cm×31cm), ensuring that the screen's field of vision covered the mouse's binocular visual areas. A full-screen drifting sinusoidal grating was presented on the screen as a visual stimulus source using visual stimulus generation software (the open-source toolkit PsychoPy in Python).

[0022] ③ Subsequently, a standardized stimulation and acquisition paradigm was implemented. To induce specific neural oscillations, the spatial frequency of the grating was set to 0.04 cpd (cycles / degree), and the temporal frequency to 2 Hz. The stimulation sequence consisted of 12 uniformly distributed orientations from 0° to 330° (30° step). Data acquisition employed an event-related design, with each orientation's stimulus presentation lasting 2 seconds (Stimulus ON), followed by a 1-second isoluminance gray screen interval (ISI / Stimulus OFF), thus constituting a complete trial lasting 3 seconds.

[0023] ④ Finally, signal synchronization recording is performed. The stimuli from the above 12 orientations are played in a pseudo-random order and repeated 20 to 30 blocks to obtain sufficient stacking average counts. While providing visual stimulation, the electrophysiological signals of the graft neurons are simultaneously recorded using a multi-channel data acquisition system, and the raw data containing precise timestamps is transmitted to the workstation for storage as the basis for subsequent feature extraction.

[0024] Furthermore, the specific frequency band is the Low Gamma band; the extraction process includes: bandpass filtering the local field potential signal to obtain the neural oscillation signal of the Low Gamma band, and calculating the relative power change of the oscillation signal relative to the baseline during the presentation of the visual stimulus.

[0025] Specifically, the implementation process of this embodiment includes: Step S2: Electrophysiological Feature Extraction and Network Analysis Based on Low Gamma Band Specific conditions and requirements: Local field potential (LFP) preprocessing and Low Gamma band energy extraction: ① First, the acquired raw LFP signal is denoised. A 4th-order Butterworth bandpass filter (filtfilt function) is used in Matlab 2021b to perform bidirectional zero-phase filtering, specifically extracting the neural oscillation signal in the Low Gamma band (30-45Hz). Then, the Hilbert transform (hilbert function) is used to obtain the analytic signal of the filtered signal, and the square of its instantaneous amplitude is calculated to obtain the instantaneous power signal P(t).

[0026] ; in, The signal is an LFP signal after being bandpass filtered at 30-45Hz. This represents the Hilbert Transform. The imaginary unit, This is the instantaneous power envelope of that frequency band.

[0027] ② Calculation of visual evoked response intensity (ΔPower): To quantify the energy change induced by visual stimuli, baseline correction was performed using the logarithmic ratio method (Decibel, dB). The baseline window was defined as -0.4 s to -0.1 s before visual stimulus presentation, and the response window was defined as 0.3 s to 1.3 s after stimulus presentation. For each trial, the ratio of the average power of the response window to the average power of the baseline window was calculated and converted to decibels (dB). ; in, The relative response intensity of a single trial in the Low Gamma band (unit: dB). The average power of the signal within the response window (0.3s~1.3s) after the presentation of the visual stimulus; The average power of the signal within the baseline window (-0.4s to -0.1s) before the presentation of the visual stimulus.

[0028] Subsequently, the average response intensity R(θ) under that orientation is obtained by averaging all valid trials for the same stimulus orientation.

[0029] Furthermore, the electrophysiological index used to characterize the selectivity of visual features is the direction selectivity index; the calculation process includes: based on the relative power values ​​of the response of the Low Gamma band neural oscillation signal under different visual stimulus orientations, the direction selectivity index is calculated by vector synthesis method.

[0030] Specifically, the implementation process of this embodiment includes: In step S2, ③ the calculation of the Directional Selectivity Index (OSI): The OSI (Optical System Indices) was calculated using a "global vector synthesis" method to assess the graft's tuning sensitivity to different orientations. To avoid negative response values ​​interfering with the vector calculation, the average response intensity for all orientations was first non-negatively linearly shifted. The OSI calculation formula is as follows: ; Where N is the total number of stimulus orientations (12 in this example), θ k Let R(θ) be the angle (in radians) of the k-th stimulus. k ) represents the average response intensity after translation, and j is the imaginary unit. The OSI value calculated by this formula is between 0 and 1, with a higher value indicating a stronger preference of the neuron population for a specific orientation.

[0031] Furthermore, topological indicators used to characterize the degree of brain network integration include global efficiency and clustering coefficient; the calculation process includes: constructing a functional connectivity network based on low-gamma band neural oscillation signals acquired through multiple channels, and then calculating the global efficiency and clustering coefficient of the functional connectivity network.

[0032] Specifically, the implementation process of this embodiment includes: In step S2, ④ Network construction and topology analysis: Using the "filtfilt" function in Matlab 2021b, the Low Gamma band (30-45Hz) signal is extracted from multi-channel LFP data. The "synchronous likelihood estimation" algorithm (HERME open-source toolbox in Matlab) is used to calculate the nonlinear functional connectivity strength between channels and construct a weighted network matrix. ; in, X k,i Indicates the first kEach channel at time i The state vector; x k,i This is the original LFP signal sequence; d This refers to the embedding dimension. This is the time delay. This step maps the one-dimensional time series to... d In a 3D state space.

[0033] ; in, SL k Indicates channel k With channel n The average synchronous likelihood intensity between them; N The total number of time points; |·| represents the Euclidean distance; r k and r n This is the critical distance threshold; θ This is the Heaviside step function (1 when the distance is less than a threshold, 0 otherwise). This formula quantifies the probability that two channels simultaneously exhibit "recursive" behavior in the state space.

[0034] Furthermore, the "clustering_and_efficiency" script function, written in Matlab 2021b, was used to calculate the network's "global efficiency" and "clustering coefficient" to quantitatively evaluate the orderliness of local microcircuits in the eye-brain assembly and its functional integration with the host whole-brain network. Clustering coefficient C : Assuming there are nodes Define the clustering coefficient of this node. The number of edges between adjacent nodes and the nodes actually connected to it. and the maximum possible number of connected edges The ratio, that is: ; in, Represents nodes in the network The actual number of connected edges, Represents nodes The number of adjacent nodes.

[0035] Then the clustering coefficient of the network C The average of the clustering coefficients for all n nodes is: ; in, nIndicates the number of nodes. ,when C When =1, the network is fully connected; C When = 0, all nodes in the network are isolated nodes with no connecting edges.

[0036] Overall efficiency E : Overall efficiency E Defined as the average of the reciprocals of the shortest path lengths between all nodes in the network, it is calculated as follows: ; in, Represents any node and The shortest path length between them. n Indicates the number of nodes.

[0037] The overall signal processing flow is as follows Figure 2 As shown.

[0038] The purpose of this step is as follows: OSI is used to determine whether the repaired neurons possess the advanced function of "feature extraction" (orderly growth vs. disordered growth); Low Gamma energy reflects the activity and synchronicity of local neural microcircuits; network properties reflect the degree of integration between the graft and the host brain.

[0039] Furthermore, the hierarchical visual behavioral tests include: the light and dark box test for testing basic light perception ability, the visual cliff test for testing depth perception ability, and the visual water maze test for testing graphic spatial discrimination ability.

[0040] Specifically, the implementation process of this embodiment includes: Step S3: Hierarchical visual behavioral function verification Specific conditions and requirements: The following three types of experiments will be conducted in order of increasing complexity of visual cognition: ① Basic light perception test: A homemade light / dark box was used for mice. The left side was the light box, and the right side was the dark box. The mice were initially placed in the light box, and their behavior was recorded using an infrared camera. The recorded metric was the number of times the mice moved between boxes.

[0041] ② Depth perception test: A homemade visual cliff device for mice was used, with one half placed on a table and the other half suspended in the air. Mice with good vision were more likely to stay on the table side than the cliff side. The recorded indicators were the proportion of time the mice spent on the "cliff side" and the number of times they moved between the two sides.

[0042] ③ Spatial Vision Test: A visual water maze was used for mice. Mice were placed at the initial position in the maze and needed to swim to the target side for the safety island to open, allowing the mice to rest. Mice with better vision showed a higher accuracy rate. Mice were required to find the safety platform based on a pattern of sinusoidal gratings, and the accuracy rate was recorded.

[0043] The purpose of this step is to verify the biological significance of electrophysiological signals from a macroscopic perspective. Level 1 verifies whether retinal signals are transmitted to the brain (excluding total blindness); Levels 2 & 3 specifically verify the functional recovery of the V1 visual cortex (excluding interference from subcortical reflex pathways), directly corresponding to the OSI enhancement in step S2.

[0044] Furthermore, the process of binarizing the electrophysiological indicators includes: setting judgment thresholds for the directional selectivity index, the global efficiency, and the clustering coefficient, counting the number of indicators that reach the corresponding thresholds, and determining that the electrophysiological dimension score meets the standard when the number of indicators that reach the thresholds is not less than the set value.

[0045] Specifically, the implementation process of this embodiment includes: Step S4: Multimodal joint evaluation based on "binary decision matrix" Specific conditions and requirements: ① Construction and standardization of feature vectors: First, extract six key features from two dimensions in steps S2 and S3 to construct a multimodal feature set: Electrophysiological feature set ( F ephys ): Includes 3 indicators. 1. Low Gamma OSI ( OSI L-γ ): Characterizes the directional selectivity of local microloops; 2. Global efficiency ( E glob ): Characterizes the degree of integration of the whole brain network; 3. Clustering coefficient ( C ): Characterizes the tightness of local network connections.

[0046] Behavioral feature set ( F behav ): Includes 3 indicators. 1. Percentage of time spent in the "open" and "closed" boxes (the ratio of time spent in the "open" box to the time spent in the "closed" box). R light ): Represents basic light and dark perception ability; 2. Proportion of time spent on the "cliffside" of the visual cliff ( R cliff ): Represents depth perception; 3. Visual water maze accuracy ( Acc maze ): Characterizes the spatial resolution of a graphic.

[0047] ② Dimensional scoring and binarization determination (threshold setting): To transform complex multidimensional data into a decidable logical state, the following scoring model is established: X-axis: Binarization of the electrophysiological dimension ( X state Scoring rules: Set effective thresholds for each feature ( OSI L-γ >0.3, E glob >0.4, C >0.3). 1 point is awarded for each fulfilled condition, and the total score is recorded as follows: Score ephys (3 points maximum). Judgment logic: Set the dimension threshold as follows: Score ephys ≥ 2. If Score ephys ≥ 2, judge X state = 1 (representing significant repair of microscopic functions); if Score ephys <2, judge X state = 0 (represents unrepaired or disordered micro-functions). Y-axis: Binarization of behavioral dimensions ( Y state Scoring rules: Set effective thresholds for each feature. Acc maze >70% (significantly higher than the random level); R cliff <30% (showing significant visual cliff phobia); R light <35% (exhibiting significant light-avoidance behavior). 1 point is awarded for each criterion met, with a total score of [total score not specified]. Score behav (3 points maximum). Judgment logic: Set the dimension threshold as follows: Score behav ≥ 2. If Score behav ≥ 2, judge Y state = 1 (representing effective visual perception); if Score behav <2, judge Y state = 0 (representing no improvement in behavioral performance).

[0048] Furthermore, the joint decision matrix is ​​a two-dimensional matrix, where one dimension represents whether the electrophysiological index score meets the standard, and the other dimension represents whether the behavioral index score meets the standard; the four-category assessment conclusion includes: both electrophysiological and behavioral scores meet the standard, corresponding to effective functional repair; only the electrophysiological score meets the standard, corresponding to invalid connectivity; only the behavioral score meets the standard, corresponding to non-cortical-dependent compensation; and neither meets the standard, corresponding to no repair.

[0049] Specifically, the implementation process of this embodiment includes: In step S4, ③ construct 2 2. Joint evaluation matrix: by X state The horizontal axis is... Y state Using the vertical axis as the ordinate, construct a decision matrix. The system bases its decisions on the final landing point (...). X state , Y state Output evaluation conclusion: Quadrant 1 (1, 1): Effective Repair. Criteria: X state = 1 and Y state = 1. Conclusion: The graft not only formed an ordered micro-network, but also successfully transformed into macroscopic visual behavioral output. Quadrant 4 (1, 0): Ineffective Connection. Judgment Criteria: X state = 1 and Y state = 0. Conclusion: Local reconstruction was successful, but an effective long-range projection or motor output pathway was not established. Second quadrant (0, 1): Non-cortical compensation. Judgment criteria: X state = 0 and Y state = 1. Conclusion: Functional recovery may originate from subcortical pathways such as the superior colliculus (SC), rather than V1 cortical repair. Third quadrant (0, 0): No Repair. Judgment criteria: X state = 0 and Y state = 0. Conclusion: The transplant failed.

[0050] Figure 1This is a visual stimulus paradigm diagram; the paradigm employs an alternating "ON-OFF" blocking design. Each complete trial (One Trial) consists of a 2-second visual stimulus presentation phase (Stim, On-period) and a 1-second gray screen interval phase (Interval, Off-period). During the Stim phase, drifting sinusoidal gratings with different orientations (0°-330°) are presented randomly. During the Interval phase, a gray screen of equal brightness is presented to eliminate visual afterimages and serve as a baseline.

[0051] Figure 2 This is a flowchart illustrating the joint assessment of eye-brain assembly functional repair based on multimodal data. The flowchart demonstrates a hierarchical assessment strategy from microscopic mechanism exploration to macroscopic functional verification. First, at the microscopic level (S1-S2), LFP signals in the brain are collected, and key electrophysiological features in the Low Gamma band (such as direction-selective OSI and network global efficiency) are extracted. Then, at the macroscopic level (S3), visual function performance is verified using a hierarchical behavioral paradigm. Finally, a binary decision matrix (S4) is used to jointly determine the microscopic and macroscopic data, outputting a four-category assessment conclusion including "effective functional repair."

[0052] Figure 3 This is a flowchart of signal processing and feature extraction based on the Low Gamma band. The flowchart illustrates the computational path from the raw LFP signal to key electrophysiological features. First, the signal is preprocessed and filtered in the Low Gamma band (30-45Hz). Then, it is divided into two parallel paths: Branch A uses Hilbert transform and vector synthesis algorithms to calculate the orientation selectivity index (OSI); Branch B uses synchronous likelihood estimation (SL) to construct a brain functional network and calculate global efficiency and clustering coefficients. Finally, the aforementioned features are output as the evaluation criteria.

[0053] In this embodiment, a computer terminal device is provided, including: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the multimodal joint evaluation method for the repair effect of the eye-brain assembly described above.

[0054] In this embodiment, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the multimodal joint evaluation method for the repair effect of the eye-brain assembly described above.

[0055] This invention provides a multimodal joint evaluation method for the repair effect of the eye-brain assembly. By integrating multimodal data from microscopic electrophysiology and macroscopic behavioral science, it achieves a multidimensional quantitative characterization of the repair effect of the visual cortex. This invention combines high spatiotemporal resolution extracellular electrophysiological signals with hierarchical visual behavioral indicators to construct a complete evaluation link from microscopic neural circuits to macroscopic visual perception, overcoming the limitations of single evaluation methods. By establishing a core evaluation index based on the Low Gamma band directional selectivity index, it can specifically identify whether neurons have formed an ordered functional network that finely processes visual features, significantly improving the specificity and sensitivity of the evaluation. Simultaneously, functional brain network analysis is introduced to quantitatively reveal the degree of integration between the graft and the host visual network. Finally, by constructing a binary joint decision matrix of microscopic and macroscopic data for cross-validation and logical judgment, interference factors such as non-cortical-dependent compensation and invalid connections are effectively eliminated. This establishes a standardized, quantitative, and definitive method for evaluating the functional repair of the eye-brain assembly, significantly improving the accuracy, robustness, and universality of the evaluation.

[0056] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multimodal joint evaluation method for the repair effect of the eye-brain assembly, characterized in that, include: Visual stimuli with different orientations were presented to experimental subjects who had received eye-brain assemblies, and local field potential signals of their primary visual cortex were collected simultaneously. The local field potential signal is processed to extract neural oscillation signals of a specific frequency band, and based on the oscillation signals, electrophysiological indicators for characterizing the selectivity of visual features and topological indicators for characterizing the degree of brain network integration are calculated. The experimental subjects underwent hierarchical visual behavioral tests, from simple to complex, to obtain behavioral indicators corresponding to different levels of visual perception. Based on preset thresholds, the electrophysiological and behavioral indicators are binarized and scored respectively. A joint decision matrix is ​​constructed based on the scoring results, and a four-category evaluation conclusion on the repair effect of the eye-brain assembly is output.

2. The method according to claim 1, characterized in that, The visual stimulus is a drifting sinusoidal grating presented on the screen. The presentation paradigm of the visual stimulus is an event-related design. Each trial includes a visual stimulus presentation period and a gray screen interval period with equal brightness. The stimuli from different directions are presented in a pseudo-random order in multiple cycles.

3. The method according to claim 2, characterized in that, The specific frequency band is the Low Gamma band; the extraction process includes: bandpass filtering the local field potential signal to obtain the neural oscillation signal of the Low Gamma band, and calculating the relative power change of the oscillation signal relative to the baseline during the presentation of the visual stimulus.

4. The method according to claim 3, characterized in that, The electrophysiological index used to characterize the selectivity of visual features is the direction selectivity index. The calculation process includes: based on the relative power values ​​of the response of the Low Gamma band neural oscillation signal under different visual stimulus orientations, the direction selectivity index is calculated by vector synthesis.

5. The method according to claim 3, characterized in that, Topological metrics used to characterize the degree of integration in brain networks include global efficiency and clustering coefficient; The calculation process includes: constructing a functional connectivity network based on low-gamma band neural oscillation signals acquired through multiple channels, using a synchronous likelihood estimation algorithm, and then calculating the global efficiency and clustering coefficient of the functional connectivity network.

6. The method according to claim 1, characterized in that, The hierarchical visual behavioral tests include: the light and dark box test for testing basic light perception, the visual cliff test for testing depth perception, and the visual water maze test for testing graphic spatial discrimination.

7. The method according to claim 5, characterized in that, The process of binary scoring of electrophysiological indicators includes: setting judgment thresholds for the directional selectivity index, the global efficiency and the clustering coefficient respectively; counting the number of indicators that reach the corresponding threshold; and determining that the electrophysiological dimension score meets the standard when the number of indicators that reach the threshold is not less than the set value.

8. The method according to claim 1, characterized in that, The joint decision matrix is ​​a two-dimensional matrix, where one dimension represents whether the electrophysiological index score meets the standard and the other dimension represents whether the behavioral index score meets the standard. The four-category assessment conclusion includes: both electrophysiological and behavioral scores meet the standard, which corresponds to effective functional repair; only the electrophysiological score meets the standard, which corresponds to invalid connectivity; only the behavioral score meets the standard, which corresponds to non-cortical-dependent compensation; and neither meets the standard, which corresponds to no repair.

9. A computer terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors perform the steps of the method as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-8.

Images