Adaptive closed-loop training method for brain optic nerve remodeling
By identifying users' visual symptoms and generating balance training programs, this approach solves the problems of poor compliance and unquantifiable effects in traditional visual health interventions, achieving safe and precise visual function training results.
Patent Information
- Application Number
- CN202511256141.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-04
- Publication Date
- 2025-11-11
AI Technical Summary
Traditional visual health intervention methods rely on professional guidance from operators, resulting in poor user compliance, unquantifiable training effects, and unsafe and inefficient training processes when multiple visual functional defects exist.
By collecting multi-dimensional visual function data from users, the primary target symptoms and secondary constraint symptoms are identified, and a balanced training plan is generated, including parameter correction and module interleaving. Combined with gamified training tasks, the training plan is adjusted in real time to avoid conflicts.
It enables safe and precise intervention for various visual functional defects, improves the reliability and effectiveness of personalized training, and ensures the safety and efficiency of the training process.
Smart Images
Figure CN120918925A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of visual health technology, and in particular to an adaptive closed-loop training method for brain visual neural remodeling. Background Technology
[0002] Vision is the most important channel for humans to perceive information from the outside world, and good visual function is the cornerstone for ensuring learning efficiency, quality of life, and even future career development. In recent years, visual health problems among children and adolescents worldwide have become increasingly serious, especially the trend of myopia occurring at younger ages and with a high incidence rate, which has become a public health concern.
[0003] Traditional visual health interventions have long focused primarily on correcting refractive errors, i.e., compensating for visual impairments through methods such as wearing glasses or orthokeratology lenses. However, with the deepening of research in optometry and neuroscience, the industry has gradually recognized that many visual problems, including the development and progression of myopia, eye strain, reading difficulties, and inattention, are not rooted solely in the optical structure of the eyeball. The deeper cause lies in the brain's ability to process visual information—that is, defects in brain visual function. Brain visual function is a complex system encompassing multiple dimensions, including eye position control, accommodation, convergence, binocular fusion, and eye movements. Abnormalities in any of these dimensions can disrupt the stability and comfort of the entire visual system.
[0004] To address this, a method using flip-over paddles for sensitivity training has emerged. However, this training process heavily relies on the operator's professional guidance and the user's self-discipline, making standardization and large-scale promotion difficult. Secondly, the training process is tedious and monotonous, leading to poor compliance among children. Finally, the training effect cannot be precisely quantified or tracked in real time, hindering effective feedback and adjustments. Furthermore, ensuring the safety and efficiency of the training process is crucial when users simultaneously possess multiple visual functional deficiencies, even those conflicting in their training principles. Summary of the Invention
[0005] To address the aforementioned issues, this application provides a highly reliable adaptive closed-loop training method for brain visual neural remodeling.
[0006] To achieve the above objectives, the adaptive closed-loop training method for brain visual neural remodeling designed in this application includes the following steps: S101: Collect multi-dimensional visual function data of the user, and identify and generate a user visual symptom profile containing at least one primary target symptom and one secondary constraint symptom based on the data. S102: Query the preset clinical knowledge base to determine whether the standard training program for the primary target symptom will have a negative impact on the secondary constraint symptom, so as to determine whether there is a training conflict; wherein, the primary target symptom is defined as a functional defect that affects the user's subjective visual perception or learning efficiency, and the secondary constraint symptom is defined as a potential or compensatory physiological state that is improperly stimulated or aggravated during the training process. S103: When a training conflict is identified, a balanced training scheme shall be generated using at least one of the following methods: Parameter correction: Select a benchmark training module for the primary target symptom, and correct at least one training parameter in the benchmark training module according to the constraint rules associated with the secondary constraint symptom. Module interleaving: In the baseline training module targeting the primary target symptom, a compensatory training submodule for stabilizing the secondary constraint symptom is interleaved. S104: Guide users to perform gamified training tasks based on the balance training scheme and collect their training performance data; S105: After one training cycle, based on the newly acquired visual function data and the training performance data, repeat steps S101 to S103 to adjust the balance training scheme.
[0007] Preferably, the primary target symptom is determined to be convergence insufficiency, which is based on detecting that the user's convergence near point value is greater than 10 cm and the positive fusion range is less than 15 prism diopters; the secondary constraint symptom is determined to be latent exotropia, which is based on measuring an outward deviation of the eye position through a cover-to-cover test.
[0008] Preferably, the acquisition of multi-dimensional visual function data in step S101 specifically includes performing the following measurement operations: measuring the adjustment sensitivity using an automatic flip-type camera, in units of cycles per minute; measuring the range of stepwise fusion convergence and divergence using a prism rod, in units of prism diopters; and measuring the lag or lead of the adjustment response using a MEM dynamic retinoscope, in units of diopter.
[0009] Preferably, determining the existence of a training conflict in step S102 specifically includes: the standard training scheme for insufficient fusion is an active depletion of the user's physiological reserves of fusion fusion ability; while in order to keep latent exotropia from manifesting, the user needs to continuously deplete the physiological reserves of the same fusion fusion ability; when it is determined that the sum of the active depletion and the basic depletion exceeds the user's safe reserve limit, a training conflict is determined to exist.
[0010] Preferably, the parameter correction is implemented as follows: in a training game simulating Brok's string, the breakpoint value of the positive fusion range measured by the user is retrieved; then the maximum horizontal separation of the two disparity images that the user is required to fuse in the game is set to a prism power requirement equivalent to 75% of the breakpoint value.
[0011] Preferably, the specific implementation of the module interleaving is as follows: after a training module that requires the user to complete 20 consecutive rapid zoom adjustments between virtual near and far targets within 2 minutes, a 1-minute window-viewing compensation sub-module is entered. This sub-module only presents a static landscape image with the focus at infinity to guide the user to relax the ciliary muscle.
[0012] Preferably, the specific process of collecting training performance data in step S104 includes: in a game that requires the user's eyeballs to smoothly follow a randomly moving light spot, the root mean square error between the user's eyeball movement trajectory and the light spot target trajectory is recorded and calculated in real time using a built-in infrared eye-tracking camera; and after the task is completed, the user is asked to mark the degree of visual fatigue they currently feel on a numerical rating scale from 0 to 10.
[0013] Preferably, in step S105, before each training session begins, the user's phoria is detected. If the detected exophoria is more than 2 prism diopters higher than the initial baseline value when the user's profile was created, the original planned balanced training scheme that includes set requirements is terminated and replaced with a basic stable scheme that only includes binocular isoopia and suppression elimination training.
[0014] Preferably, in step S105, if the convergence near point value improves by more than 1 cm and the exophoria remains stable, the prism power required for convergence in the training scheme is increased by 0.5 prism powers; if the convergence near point value improves by more than 1 cm but the exophoria increases, the difficulty of the current training scheme remains unchanged, and the duration of the window-viewing compensation submodule is increased.
[0015] Preferably, the gamified training task is implemented on a screen using color separation technology, wherein the images provided to the left and right eyes are rendered in red and green respectively, the user wears corresponding red-green complementary glasses to view them, and by changing the horizontal spacing of the red and green images, the bottom-out or bottom-in prism effect required for training the fusion set or divergence function is generated.
[0016] The adaptive closed-loop training method for brain visual neural remodeling designed in this application identifies the user's primary target symptoms and secondary constraint symptoms and determines the potential training conflict between them before generating a training plan. Then, it generates a balanced training plan with parameter correction or module interleaving. This plan can improve the primary symptoms in a targeted manner while stabilizing the secondary symptoms. It can also be dynamically adjusted based on feedback after the training cycle, thereby achieving safe and accurate intervention for complex visual problems with multiple functional defects and improving the reliability of personalized training. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating an adaptive closed-loop training method for brain visual neural remodeling, provided as an embodiment of this application. Detailed Implementation
[0018] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application.
[0019] The methods described in this application embodiment can be deployed and executed in a specially designed brain visual function assessment and training system. This system is typically installed in the visual health rooms of optometry centers, hospital ophthalmology departments, professional rehabilitation institutions, or schools.
[0020] The system typically includes a user terminal in terms of hardware, such as an all-in-one computer integrating a display screen, processor, and built-in infrared eye-tracking module. This user terminal also provides standardized data interfaces for connecting to various external professional optometric measurement devices, such as automatic flip-type cameras, prism rods, and MEM dynamic retinoscopes. On the software side, the user terminal runs an application software suite that integrates a data acquisition module, a symptom profile generation module, a clinical knowledge base query module, a balance training prescription generation module, and a gamified training engine.
[0021] In the following description, "system" refers to the collaborative technical entity comprised of the user terminal and core application software. The specific process by which this system executes the method of this application is as follows: like Figure 1 As shown in this embodiment, an adaptive closed-loop training method for brain visual neural remodeling is described. This method may specifically include the following steps: S101: Generate a visual symptom profile. In step S101, the system first collects multi-dimensional visual function data from the user. In a specific application scenario, the collection operation may include: measuring the adjustment sensitivity using an automatic flip-type imager connected to the user terminal, with the result expressed in cycles per minute (cpm); measuring the range of stepwise fusion convergence and divergence using a prism rod, with the result expressed in prism diopters (PD); and measuring the hysteresis or antegrade amount of the accommodation response using a MEM dynamic retinoscope, with the result expressed in diopters (S104).
[0022] After data collection, the system's symptom profiling module identifies and generates a user visual symptom profile based on this quantified data. This profile is not a list of single symptoms, but rather includes one or more primary target symptoms and one or more secondary constraint symptoms. Primary target symptoms are defined as the functional deficiencies that currently most significantly impact the user's subjective visual experience or learning efficiency, while secondary constraint symptoms are defined as potential or compensatory physiological states that may be improperly triggered or worsened during subsequent training.
[0023] For example, taking a user with both convergence insufficiency and latent exotropia as an example, the system identifies convergence insufficiency as the primary target symptom based on the detection that the user's convergence near point (NPC) value is greater than 10 cm and its positive fusion range is less than 15 prism diopters; at the same time, based on the outward deviation of its eye position measured by the occlusion-unocclusion test or Howell card test, latent exotropia is identified as a secondary constraint symptom in the training process.
[0024] S102: Conflict Detection and Solution Arbitration. In step S102, the system queries a pre-set database containing a large amount of optometric clinical knowledge to determine whether the standard training scheme for the primary target symptoms identified in step S101 will negatively affect the secondary constraint symptoms, thereby determining whether a training conflict exists between the two. In this embodiment, the clinical knowledge base is not an information list, but a structured set of expert rules and data models.
[0025] The primary target symptom is defined as a functional defect that affects the user's subjective visual experience or learning efficiency, while the secondary constraint symptom is defined as a potential or compensatory physiological state that is improperly activated or aggravated during training. Taking the aforementioned scenario as an example, the system's conflict determination logic for this specific scenario is as follows: First, the system learns from the knowledge base that the core mechanism of the standard training scheme for insufficient fusion is to guide the user's eyes to continuously converge inward, thereby exercising the strength and neural control of the relevant intraocular and extraocular muscles. This can be regarded as an active depletion of the user's physiological reserve of fusion convergence ability. Second, the system also learns from the knowledge base that in order to maintain latent exotropia in daily life and avoid diplopia, the user needs to continuously use the same physiological reserve of fusion convergence ability for continuous basic depletion to counteract the natural tendency of the eyes to deviate outward. Finally, when the system determines that the sum of the active depletion and the basic depletion is highly likely to exceed the user's safe reserve limit, a training conflict is identified.
[0026] S103: Generate a balanced training scheme. When the system determines a training conflict in step S102, in step S103, the system will not use the standard training scheme, but will instead generate a balanced training scheme in at least one of the following ways: Preferably, the balanced training scheme can be generated through parameter modification: a benchmark training module targeting the primary target symptom is selected, and at least one training parameter in the benchmark training module is modified according to the constraint rules associated with the secondary constraint symptom. Specifically, in a 3D training game simulating a Brock string, the system first retrieves the breakpoint value (e.g., 15 PD) of the user's positive fusion range from the user's visual function data. Then, the system sets the maximum horizontal separation of the two disparity images that the user is required to fuse in the game to a prism power requirement equivalent to 75% of this breakpoint value (i.e., 11.25 PD), thereby ensuring that the training intensity effectively stimulates ensemble function while remaining within the user's safe compensatory capacity.
[0027] Optionally, this balance training program can also be generated through module interleaving: within a baseline training module targeting the primary objective symptoms, a compensatory training sub-module for stabilizing the secondary constraint symptoms is interspersed. Specifically, after a training module requiring the user to continuously complete 20 virtual near-far targets within 2 minutes, such as simulating rapid zooming between 40 cm and 4 meters, the system forces the user into a 1-minute compensatory sub-module of viewing distant objects. This sub-module presents only a static landscape image with the focus at infinity, actively guiding the user's ciliary muscles to achieve complete relaxation, thereby offsetting the risk of accommodative spasm that may have arisen from the previous high-intensity accommodative training.
[0028] S104: Execution and Feedback. In step S104, the user executes a series of gamified training tasks according to the balance training scheme generated in step S103, while the system collects performance data during the training process.
[0029] In a preferred embodiment, the gamified training task is implemented on a screen using color separation technology. Images provided to the left and right eyes are rendered in red and green, respectively. The user must wear corresponding red-green complementary glasses to view the images. The system precisely alters the horizontal spacing between the red and green images to generate the bottom-out or bottom-in prism effect required for training fusion or divergence functions, thereby achieving targeted training of binocular vision.
[0030] During the training process, the specific data collection process may include: in a game that requires the user's eyes to smoothly follow a randomly moving light spot, the built-in infrared eye-tracking camera records and calculates the root mean square error between the user's eye movement trajectory and the target light spot trajectory in real time, as an objective performance indicator; after the task is completed, the system will pop up a graphical interface, requiring the user to drag a slider on a numerical scale from 0 to 10 to indicate the degree of visual fatigue they currently feel, as a subjective feedback indicator.
[0031] S105: Closed-loop adjustment. In step S105, the system adaptively and dynamically adjusts the balance training scheme based on the newly collected visual function data and the training performance data within the period after a preset training cycle, such as one week.
[0032] The dynamic adjustment mechanism preferably includes a safety coverage logic. Specifically, before each training session begins, the system first performs a rapid detection of the user's exophoria. If the detected exophoria is more than two prism diopters higher than the initial baseline value when the user's profile was created, the system determines that the user's compensatory ability has decreased. It then immediately terminates the original planned balanced training program that includes set requirements and forcibly replaces it with a basic stable program that only includes binocular isoopia and suppression elimination training to prevent risk accumulation.
[0033] Without triggering the safety coverage logic, dynamic adjustments can be executed based on a well-defined rule-based decision tree. For example, this decision tree might include the following rules: if the system detects that the user's ensemble near point has improved by more than 1 cm and the exophoria remains stable, the system increases the prism strength of the ensemble requirement in the training scheme by 0.5 prisms to enhance training intensity; if the system detects that the ensemble near point has improved by more than 1 cm but the exophoria has increased, the system maintains the current training scheme's difficulty level and increases the duration of the window-gazing compensation submodule to strengthen the relaxation phase.
[0034] This application provides an adaptive closed-loop training method for brain visual neural remodeling. Before generating a training plan, the method first identifies the user's primary target symptoms and secondary constraint symptoms and determines the potential training conflict between them. Then, it generates a balanced training plan with parameter correction or module interleaving. This plan can improve the primary symptoms in a targeted manner while stabilizing the secondary symptoms. It can also be dynamically adjusted based on feedback after the training cycle, thereby achieving safe and accurate intervention for complex visual problems with multiple functional defects and improving the reliability of personalized training.
[0035] In the description of this application, it should be noted that the terms vertical, up, down, horizontal, etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.
[0036] In the description of this application, it should also be noted that, unless otherwise explicitly specified and limited, the terms "setup," "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.
[0037] Finally, it should be noted that the above descriptions are merely preferred embodiments of this application and are not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
Claims
1. An adaptive closed-loop training method for brain visual neural remodeling, characterized in that, Includes the following steps: S101: Collect multi-dimensional visual function data of the user, and identify and generate a user visual symptom profile containing at least one primary target symptom and one secondary constraint symptom based on the data. S102: Query the preset clinical knowledge base to determine whether the standard training program for the primary target symptom will have a negative impact on the secondary constraint symptom, so as to determine whether there is a training conflict; wherein, the primary target symptom is defined as a functional defect that affects the user's subjective visual perception or learning efficiency, and the secondary constraint symptom is defined as a potential or compensatory physiological state that is improperly stimulated or aggravated during the training process. S103: When a training conflict is identified, a balanced training scheme shall be generated using at least one of the following methods: Parameter correction: Select a benchmark training module for the primary target symptom, and correct at least one training parameter in the benchmark training module according to the constraint rules associated with the secondary constraint symptom. Module interleaving: In the baseline training module targeting the primary target symptom, a compensatory training submodule for stabilizing the secondary constraint symptom is interleaved. S104: Guide users to perform gamified training tasks based on the balance training scheme and collect their training performance data; S105: After one training cycle, based on the newly acquired visual function data and the training performance data, repeat steps S101 to S103 to adjust the balance training scheme.
2. The adaptive closed-loop training method for brain visual neural remodeling according to claim 1, characterized in that, The primary target symptom was identified as convergence insufficiency, based on the detection that the user's convergence near point was greater than 10 cm and the positive fusion range was less than 15 prism diopters; the secondary constraint symptom was identified as latent exotropia, based on the outward deviation of the eye position measured by occlusion-to-occlusion testing.
3. The adaptive closed-loop training method for brain visual neural remodeling according to claim 2, characterized in that, The acquisition of multi-dimensional visual function data in step S101 specifically includes performing the following measurement operations: measuring the adjustment sensitivity using an automatic flip-type camera, in units of cycles per minute; measuring the range of stepwise fusion convergence and divergence using a prism rod, in units of prism diopters; and measuring the lag or lead of the adjustment response using a MEM dynamic retinoscope, in units of diopter.
4. The adaptive closed-loop training method for brain visual neural remodeling according to claim 3, characterized in that, The determination of a training conflict in step S102 specifically includes: the standard training scheme for insufficient fusion is an active depletion of the user's physiological reserves of fusion fusion ability; while in order to keep latent exotropia from appearing, the user needs to continuously deplete the physiological reserves of the same fusion fusion ability; when it is determined that the sum of the active depletion and the basic depletion exceeds the user's safe reserve limit, a training conflict is determined to exist.
5. The adaptive closed-loop training method for brain visual neural remodeling according to claim 4, characterized in that, The specific implementation of the parameter correction is as follows: in a training game simulating Brok's string, the breakpoint value of the positive fusion range measured by the user is retrieved; then the maximum horizontal separation of the two disparity images that the user is required to fuse in the game is set to a prism power requirement equivalent to 75% of the breakpoint value.
6. The adaptive closed-loop training method for brain visual neural remodeling according to claim 5, characterized in that, The specific implementation of the interleaved modules is as follows: After a training module that requires the user to complete 20 rapid zoom adjustments between virtual near and far targets within 2 minutes, a 1-minute window-viewing compensation sub-module is entered. This sub-module presents only a static landscape image with the focus at infinity to guide the user to relax their ciliary muscles.
7. The adaptive closed-loop training method for brain visual neural remodeling according to claim 6, characterized in that, The specific process of collecting training performance data in step S104 includes: in a game that requires the user's eyeballs to smoothly follow a randomly moving light spot, the root mean square error between the user's eyeball movement trajectory and the light spot target trajectory is recorded and calculated in real time through a built-in infrared eye-tracking camera; and after the task is completed, the user is asked to mark the degree of visual fatigue they currently feel on a numerical rating scale from 0 to 10.
8. The adaptive closed-loop training method for brain visual neural remodeling according to claim 7, characterized in that, In step S105, before each training session begins, the user's phoria is detected. If the detected exophoria is more than 2 prism diopters higher than the initial baseline value when the user's profile was created, the original planned balanced training scheme that includes set requirements is terminated and replaced with a basic stable scheme that only includes binocular isoopia and suppression elimination training.
9. The adaptive closed-loop training method for brain visual neural remodeling according to claim 8, characterized in that, In step S105, if the convergence near point value improves by more than 1 cm and the exophoria remains stable, the prism power required for convergence in the training scheme is increased by 0.5 prism powers; if the convergence near point value improves by more than 1 cm but the exophoria increases, the difficulty of the current training scheme remains unchanged, and the duration of the window-viewing compensation submodule is increased.
10. The adaptive closed-loop training method for brain visual neural remodeling according to claim 9, characterized in that, The gamified training task is implemented on a screen using color separation technology, where images provided to the left and right eyes are rendered in red and green respectively. The user views the images while wearing corresponding red-green complementary glasses, and the bottom-out or bottom-in prism effect required for training the fusion set or divergence function is generated by changing the horizontal spacing of the red and green images.