Wireless photovoltaic visual cortex prosthesis system and method for total blind visual reconstruction

Abstract

The invention discloses a wireless photovoltaic visual cortex prosthesis system and method for total-blindness visual reconstruction, and the system comprises an external image processing unit which is used for collecting an environment image and generating a corresponding optical stimulation pattern; the near-infrared projection unit is connected with the external image processing unit and is used for converting the optical stimulation pattern into a near-infrared light stimulation sequence; the optical transparent cranial window assembly is used for transmitting the near-infrared light stimulation sequence to the visual cortex; and the flexible photovoltaic stimulation array is used for receiving the near-infrared light stimulation sequence through the optical transparent cranial window assembly and directly converting the near-infrared light stimulation sequence into an electrical stimulation signal for stimulating the primary visual cortex V1. By combining the flexible photovoltaic stimulation array with the optical transparent cranial window assembly, a wireless photoelectric link from the outside to the primary visual cortex is constructed, and a solution for reconstructing functional vision by directly stimulating the cortex without a leather-covered cable is provided for all-blind patients.

Description

Technical Field

[0001] This invention relates to the field of total blindness vision reconstruction technology, and particularly to a wireless photovoltaic visual cortex prosthesis system and method for total blindness vision reconstruction. Background Technology

[0002] Currently, visual prosthesis technology, as an important research direction for restoring visual function in the blind, has made significant progress in recent years. Existing visual reconstruction solutions mainly fall into two categories: retinal prostheses and visual cortical stimulation systems. Retinal prostheses, such as the PRIMA system, capture environmental images through a glasses-like camera device. After optimization by a processor, the image is converted into light signals by a near-infrared projection unit. These signals are then received by a photovoltaic array implanted under the retina and converted into electrical stimulation, thereby activating residual retinal cells to generate visual perception. Although these systems achieve wireless operation, they are only suitable for patients with partial retinal degeneration but intact optic nerves, and cannot provide visual reconstruction for totally blind patients with complete eyeball or optic nerve damage. For totally blind patients, researchers have developed solutions that directly stimulate the primary visual cortex V1. For example, Neuralink's Blindsight system uses an electrode array directly implanted in the V1 region and transmits stimulation signals through a wire penetrating the scalp. However, this wired design significantly increases the risk of infection, and after long-term implantation, the system failure rate due to wire fatigue and battery degradation is as high as 30%. Therefore, establishing an efficient optical pathway from outside the body to the visual cortex, enabling near-infrared light to efficiently penetrate the scalp and skull to provide sufficient energy for the photovoltaic array implanted in the cortex, has become a key technical bottleneck for achieving long-term stable visual reconstruction in totally blind patients. Summary of the Invention

[0003] In view of this, the present invention proposes a wireless photovoltaic visual cortical prosthesis system and method for visual reconstruction in total blindness, which can achieve long-term stable visual reconstruction in total blind patients. The present invention provides the following technical solution: A wireless photovoltaic visual cortical prosthesis system for visual reconstruction in the blind, the system comprising: An external image processing unit is used to acquire environmental images and generate corresponding optical stimulus patterns; A near-infrared projection unit, connected to the external image processing unit, is used to convert the optical stimulation pattern into a near-infrared light stimulation sequence. An optically transparent cranial window assembly is implanted in a predetermined location in the subject's skull and includes a light-transmitting structure optically coupled to the skull, which transmits the near-infrared light stimulation sequence to the visual cortex. A flexible photovoltaic stimulation array, implanted in the functional area of ​​the primary visual cortex V1, includes multiple photovoltaic conversion units and corresponding stimulation electrodes. The flexible photovoltaic stimulation array is used to receive the near-infrared light stimulation sequence through the optically transparent cranial window assembly and convert it into an electrical stimulation signal to stimulate the primary visual cortex V1. The near-infrared projection unit and the flexible photovoltaic stimulation array are wirelessly coupled via the optically transparent cranial window.

[0004] Optionally, the flexible photovoltaic stimulation array is implanted at a depth of 1.0-2.0 mm below the dura mater of the visual cortex, and the flexible photovoltaic stimulation array is fixed to the curved surface of the visual cortex by means of a PEG-DA gel coating.

[0005] Optionally, each of the photovoltaic conversion units includes multiple photovoltaic diodes connected in series, wherein the photovoltaic diodes are made of InGaAs material.

[0006] Optionally, the electrical stimulation signal output by the photovoltaic conversion unit is a biphase charge balance pulse; in the biphase charge balance pulse, the charge density of each phase pulse does not exceed the safe reversible charge injection limit of the material used for the stimulation electrode.

[0007] Optionally, the optically transparent craniotomy assembly includes a titanium alloy fixation device and a PDMS optical film, wherein the PDMS optical film covers the inner side of the titanium alloy fixation device, and the PDMS optical film has a transmittance of more than 90% for near-infrared light with a wavelength of 850-940nm; The optically transparent craniotomy assembly also includes a scalp-adaptive optical patch covering the scalp surface, wherein the refractive index of the scalp-adaptive optical patch is designed to match the refractive index of the PDMS optical film.

[0008] Optionally, the external image processing unit includes an image optimization module, which runs an edge detection algorithm and a super-resolution reconstruction algorithm; The external image processing unit is also equipped with a pre-adjustment frame generation module, which generates a sub-threshold near-infrared light pre-adjustment frame to optically pre-charge the photovoltaic conversion unit in the flexible photovoltaic stimulation array that will darken in the next frame.

[0009] Optionally, the near-infrared light output wavelength of the near-infrared projection unit is 850-915nm, and the projection frame rate is not less than 1000fps; The near-infrared projection unit also includes a MEMS micromirror alignment component, which is used to accurately project the near-infrared light stimulation sequence onto the optically transparent cranial window component, with a projection alignment error of less than 50μm.

[0010] Optionally, the system further includes a safety monitoring module, which is used to monitor in real time the temperature change of the flexible photovoltaic stimulation array, the charge density of the electrical stimulation signal, and the output power of the near-infrared projection unit.

[0011] The present invention further discloses a method for visual reconstruction in total blindness, the method comprising: An environmental image is acquired by an external image processing unit, and the environmental image is subjected to feature optimization processing to generate an optical stimulation pattern adapted to the visual cortex perception. Based on the optical stimulation pattern, a sub-threshold near-infrared light pre-adjustment frame is generated, and the optical stimulation pattern is converted into a stimulation-type near-infrared light signal through a near-infrared projection unit. Near-infrared light pre-frame and stimulating near-infrared light signal are projected onto the optically transparent cranial window component according to a preset time sequence. After being transmitted in coordination with the optically transparent cranial window component and the scalp-adaptive optical patch, the signal is precisely incident onto the flexible photovoltaic stimulation array. After receiving near-infrared light signals, the flexible photovoltaic stimulation array converts the light energy into electrical stimulation signals through a photovoltaic conversion unit. The electrical stimulation signals are then applied to functional neurons in the primary visual cortex V1 via stimulation electrodes. The visual reconstruction effect is monitored through a closed-loop feedback mechanism, and the parameters of the optical stimulation pattern and the electrical stimulation signal are dynamically adjusted until the visual reconstruction of the preset equivalent visual acuity is achieved.

[0012] Optionally, the generation of sub-threshold near-infrared light pre-tuning frames includes: Based on the mapping relationship between the optical stimulation pattern and the target cortical electric field, the optimal light intensity distribution for pre-biasing each photovoltaic conversion unit is calculated by solving a regularized least squares optimization problem.

[0013] According to the technical solution of the present invention, by constructing a fully wireless photoelectric conversion link, an external image processing unit, a near-infrared projection unit, an implantable optically transparent cranial window component, and a flexible photovoltaic stimulation array in the V1 region are wirelessly optically coupled. The optically transparent cranial window component establishes a low-scattering wireless optical pathway from the scalp to the visual cortex, enabling the near-infrared light stimulation sequence to penetrate efficiently without being absorbed or scattered. The flexible photovoltaic stimulation array directly converts the received wireless light signal into a precise electrical stimulation signal to activate V1 neurons. Thus, in a fully wireless working mode that requires no wires or built-in power supply, functional central vision is reconstructed for totally blind patients, achieving fully wireless visual reconstruction while reducing the risk of infection. Attached Figure Description

[0014] For illustrative and not limiting purposes, the present invention will now be described in conjunction with embodiments and accompanying drawings, wherein: Figure 1This is a schematic diagram of the composition structure of the wireless photovoltaic visual cortex prosthesis system in an embodiment of the present invention; Figure 2 This is a schematic flowchart of the total blindness visual reconstruction method in an embodiment of the present invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0016] It should be noted that, where there is no conflict, the embodiments and features of the embodiments in this application can be combined with each other. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0017] refer to Figure 1 This embodiment discloses a wireless photovoltaic visual cortical prosthesis system for visual reconstruction in complete blindness. The system includes an external image processing unit 11, a near-infrared projection unit 12, an optically transparent cranial window assembly 13, and a flexible photovoltaic stimulation array 14. The near-infrared projection unit 12 and the flexible photovoltaic stimulation array 14 are wirelessly optically coupled through the optically transparent cranial window 13. The following is a detailed description: An external image processing unit 11 is used to acquire environmental images and generate corresponding optical stimulus patterns. Specifically, the external image processing unit 11 includes an image acquisition device, which in this embodiment is exemplarily a wearable glasses structure, including a high-definition wide-angle camera, a vision processing unit, and a wireless communication module. After the camera continuously acquires environmental images, the image data is transmitted to the vision processing unit (VPU) in real time. This processor is based on the ARM architecture and is equipped with image processing software optimized specifically for this system.

[0018] The image processing software follows strict timing requirements, with the total end-to-end latency controlled within 10 milliseconds. First, the software performs adaptive brightness normalization and local contrast enhancement on the original image. Next, it calls a convolutional neural network model based on the YOLOv8 framework to perform edge detection, quickly extracting contours and key object features from the image. To further enhance perceptual detail, the software then employs a super-resolution reconstruction algorithm based on the ESRGAN model to upscale the image resolution to a format more suitable for cortical mapping. After the above preprocessing, the image is converted into a light intensity distribution matrix corresponding one-to-one with the photovoltaic pixels of the implanted photovoltaic stimulation array, based on a pre-calibrated visual spatial mapping relationship obtained through functional magnetic resonance imaging (fMRI), resulting in the final optical stimulation pattern.

[0019] The external image processing unit 11 is further configured with a pre-tuned frame generation module 111. The pre-tuned frame generation module 111, based on the difference between the optical stimulus patterns of the current frame and the next frame, uses the minimum mean square error (MMSE) criterion and introduces Tikhonov regularization to prevent instability of the solution, and solves the formula... The light intensity distribution of the pre-tuned frame is calculated, where U is the photoelectric response matrix obtained by COMSOL simulation or experimental calibration, v is the target electric field vector expected to be generated in the V1 cortex, and λ is the regularization coefficient used to control the pre-tuned intensity to prevent light intensity overload. The calculated pre-tuned frame light intensity distribution x will ensure that a sub-threshold illumination is applied to the pixels in the array that will become "dark" in the next frame, causing their photovoltaic diodes to generate a bias voltage of approximately 0.3V.

[0020] The entire processing architecture maintains an end-to-end latency of less than 10ms, ensuring a real-time visual experience. The output optical stimulus pattern contains two frame types: Sub-threshold near-infrared light pre-setting frame: lasts for 10ms, light intensity 0.3mW / mm², used to precharge the photovoltaic conversion unit that will dim in the next frame to a bias voltage of 0.3V; Stimulating near-infrared light signal frame: lasts 10μs, light intensity 1.0mW / mm², generating a 20pA stimulation current.

[0021] The two outputs alternately according to a 60Hz phase-locked timing sequence and are transmitted to the near-infrared projection unit 12 via the MIPI interface.

[0022] The external image processing unit 11 also integrates a closed-loop training module 112, which collects visual evoked potentials (VEPs) through the EEG headband worn by the user, dynamically adjusts the image enhancement parameters and pre-adjustment intensity, and gradually improves the visual reconstruction quality to a 20 / 200 equivalent visual acuity level.

[0023] The near-infrared projection unit 12 is connected to the external image processing unit 11 and is used to convert the optical stimulation pattern into a near-infrared light stimulation sequence.

[0024] The near-infrared projection unit 12 employs a miniaturized VCSEL (Vertical-Cavity Surface-Emitting Laser) array module, connected to the external image processing unit 11 via a MIPI interface to receive optical stimulation pattern data streams. In this embodiment, the VCSEL array consists of 1000×1000 independently controllable laser emission points, each with a center wavelength of 850nm, precisely matching the peak response wavelength of the InGaAs photovoltaic diode in the flexible photovoltaic stimulation array 14, achieving a quantum efficiency of over 70%. The VCSEL array operates in pulse modulation mode, supporting a dual-frame working mechanism. Specifically, the pre-modulation frame phase lasts 10ms, emitting near-infrared light with a sub-threshold intensity of 0.3mW / mm², used to precharge the photovoltaic conversion unit, which will dim in the next frame, to a bias voltage of 0.3V; the stimulation frame phase lasts 10μs, emitting high-energy near-infrared light of 1.0mW / mm², generating a stimulation current of approximately 20pA in the photovoltaic conversion unit. The near-infrared projection unit 12 employs phase-locked technology to ensure precise timing coordination between the pre-set frame and the stimulation frame at a 60Hz refresh rate, with end-to-end latency controlled within 5ms. To achieve accurate spatial mapping, the near-infrared projection unit 12 integrates a MEMS (Micro-Electro-Mechanical Systems) micromirror alignment component to adjust the beam direction in real time. Combined with an eye-tracking camera to compensate for minute head movements, this ensures that the alignment error between the center of the projected light spot and the optically transparent cranial window component 13 is less than 50μm, allowing over 90% of the light energy to accurately cover the target area. The peak irradiance output by the projection unit is strictly controlled, with a maximum value of 1.0mW / mm² reaching the scalp surface. This parameter, along with the pulse width and duty cycle, is designed to ensure that the tissue temperature rise is achieved while meeting the nerve stimulation threshold. The temperature is below 1°C, and the charge density on the electrode surface is below the preset safety limit of 0.35 mC / cm² to meet the electrochemical and thermal safety requirements for long-term implantation. The near-infrared projection unit 12 is also connected to a real-time monitoring feedback loop to continuously monitor the output light intensity. When the light intensity fluctuation exceeds ±5% or the temperature rise exceeds the threshold, the drive current is automatically adjusted to maintain a safe operating state.

[0025] An optically transparent craniotomy assembly 13 is implanted at a predetermined location in the subject's skull. The optically transparent craniotomy assembly 13 includes a light-transmitting structure optically coupled to the skull for transmitting the near-infrared light stimulation sequence to the visual cortex. In this embodiment, the optically transparent craniotomy assembly 13 aims to replace a portion of the skull to construct an efficient and stable optical pathway from the scalp surface to the intracranial visual cortex. It employs a three-layer composite structure design and is implanted at a predetermined location in the occipital lobe region of the subject's skull. This location is precisely determined using preoperative fMRI and intraoperative neuronavigation systems, corresponding to the foveal center in the primary visual cortex V1. The three-layer composite structure is configured as follows: The outer layer is a medical-grade titanium alloy fixation frame (Ti-6Al-4V alloy, thickness 0.8-1.0mm), which is made into a ring structure using 3D printing technology. The outer diameter is 22mm and the inner diameter is 20mm. The edge is designed with a micro-step structure for precise engagement with the edge of the skull drill hole; The middle layer is a PDMS (polydimethylsiloxane) optical film with a thickness precisely controlled at 200μm. It is sealed and fixed to the inside of the titanium alloy frame using a hot pressing process. This PDMS film has a transmittance of more than 90% for near-infrared light with a wavelength of 850-940nm and a haze of less than 5%. It is surface encapsulated with parylene-C and alumina Al2O3 nanolayers (2-5μm) to prevent protein adhesion and material degradation; The inner layer is a scalp-adaptive optical patch made of biocompatible silicone with a thickness of 0.5mm and a refractive index precisely controlled in the range of 1.41-1.42. The refractive index is matched with that of the PDMS optical film to reduce Fresnel reflection and scattering of light at the interface. For the implantation of the optically transparent cranial window component 13, the target area V1 was first located using fMRI. Then, a circular window with a diameter of 1 cm was precisely drilled into the skull, while simultaneously thinning the scalp in the corresponding area to a thickness of 0.5 mm using a low-energy laser. The assembled optically transparent cranial window component was then embedded into the skull window and fixed to the edge of the skull step using micro-screws on a titanium alloy frame, forming a mechanical seal. Finally, a scalp-adaptive optical patch was attached to the thinned scalp surface. After implantation, optical coherence tomography (OCT) and infrared spectroscopy were used to verify the transmittance of the entire optical pathway, ensuring that the total transmittance of the optical pathway from the scalp surface to the visual cortex was greater than 90%, and the scattering loss was less than 5%. The optically transparent cranial window component 13, working in conjunction with the scalp-adaptive optical patch, forms a low-scattering optical channel, enabling the light stimulation sequence emitted by the near-infrared projection unit to efficiently penetrate the tissue, minimizing energy loss, providing sufficient light energy density to the flexible photovoltaic stimulation array beneath the cortex, while maintaining the integrity of the skull's biological barrier, reducing the long-term infection risk to below 5%, and having an expected lifespan of over 10 years.

[0026] A flexible photovoltaic stimulation array 14, implanted in the functional region of the primary visual cortex V1, includes multiple photovoltaic conversion units 141 and corresponding stimulation electrodes 142. It receives the near-infrared light stimulation sequence through the optically transparent cranial window assembly 13 and directly converts it into an electrical stimulation signal to stimulate the primary visual cortex V1. In this embodiment, the flexible photovoltaic stimulation array 14 employs an ultra-thin flexible substrate design and is precisely implanted in the functional region of the primary visual cortex V1, specifically corresponding to the 10° field of view region at the center of the foveal lobe in the posterior occipital lobe. This region is precisely located using fMRI and an intraoperative neuronavigation system to ensure that the array covers the critical area of ​​central vision in the human eye. In this embodiment, the array has an overall size of 2×2 cm and a thickness of 30 μm. A flexible silicon film serves as the carrier substrate, on which 2000 photovoltaic conversion units 141 are integrated in a honeycomb arrangement. Each unit has a diameter of 50 μm, and the center-to-center spacing between adjacent units is 50–100 µm, forming a high-density stimulation grid. Each photovoltaic conversion unit 141 contains 3-5 InGaAs photovoltaic diodes connected in series. This N-string design increases the open-circuit voltage to the range of 0.9-1.5V to overcome tissue-electrode interface resistance. When receiving 850nm near-infrared light with an intensity of 1mW / mm², it can generate a stimulation current of 15-20pA, meeting the activation threshold of approximately 10pA for V1 cortical pyramidal neurons. The photosensitive region of the photovoltaic conversion unit 141 is located on the upper surface, and the back reflective layer design improves light absorption efficiency, achieving a quantum efficiency of over 85% at 850nm wavelength. The stimulation electrode 142 is a Ptgray electrode structure, which can be a microelectrode modified with a platinum gray film or a platinum-iridium alloy electrode. It is deposited on the lower surface of the photovoltaic conversion unit 141, directly contacting the nerve tissue. Each electrode contains an active electrode and a local recirculation electrode, forming a bipolar configuration. The electrode impedance is less than 1kΩ, effectively limiting electric field diffusion and improving spatial resolution. The flexible photovoltaic stimulation array 14 is implanted at a depth of 1.5 mm (range 1.0-2.0 mm) subdurally, precisely located in the L4 / 5 stimulation layer of the cortex, where pyramidal neurons are most sensitive to electrical stimulation. The flexible photovoltaic stimulation array 14 is fixed with a biocompatible PEG-DA (polyethylene glycol diacrylate) gel coating. This coating, after UV curing, forms an elastic gel network that mechanically anchors to the cortical surface, exhibiting a displacement resistance >0.5 N / mm². In monkey model testing, displacement was less than 100 μm over 5 years. During operation, the flexible photovoltaic stimulation array 14 receives near-infrared light stimulation sequences through the optically transparent cranial window assembly 13. Each photovoltaic conversion unit directly converts the light signal into an electrical stimulation signal, employing a biphasic charge-balanced pulse pattern (cathode first, then anode, 100 μs per phase) to ensure zero net charge transfer and prevent electrode corrosion and tissue damage. The charge density of each phase is strictly controlled within the safe reversible injection limit of the Ptgray electrode, which in this embodiment is less than 0.35 millicoulombs per square centimeter.The flexible photovoltaic stimulation array 14 also integrates a pixel discharge network composed of high-resistance thin-film resistors, providing charge discharge channels during interphase and interframe periods to ensure the electrode-tissue interface is completely reset to a zero-bias state. For thermal management, it employs a 10μs short pulse and a low duty cycle (<0.1%) operating mode, combined with the high thermal conductivity (2.5W / m·K) of the flexible substrate, resulting in a tissue temperature rise ΔT <1°C. This flexible photovoltaic stimulation array 14, without requiring any wires penetrating the scalp or an internal power supply, can directly convert received light signals into precisely controlled electrical stimulation, activating V1 neurons to produce photopsia and achieving functional visual reconstruction, providing a long-term, stable visual prosthesis solution for totally blind patients.

[0027] The coordinated operation of the aforementioned external image processing unit 11, near-infrared projection unit 12, optically transparent cranial window assembly 13, and flexible photovoltaic stimulation array 14 is based on a complete optical-electrical link model and safety design principles, specifically including the optical and energy link section, the photocurrent and open-circuit voltage generation section, and the electrode-tissue electrochemistry and stimulation safety section. Among these: The optical and energy link section specifically includes: establishing an energy transfer model from the light-emitting surface of the projection unit to the dermal photovoltaic pixels. Let the peak irradiance output from the projection unit and reaching the scalp surface be denoted as... (Typical value is 1mW / mm²). Near-infrared light passes sequentially through the scalp adapter optical patch, PDMS optical film, encapsulation barrier layer, and pia mater gap, and is finally absorbed by the photovoltaic diode. The effective irradiance incident on the pixel photosensitive surface... ,in This is a correction factor introduced by the incident angle and scattering. By optimizing the materials and thickness of each layer and locally thinning the scalp during surgery, the overall transmittance of the entire process for 850nm light is greater than 90%, and the scattering loss is less than 5%. (Pixel received optical power) ,in The effective photosensitive area of ​​a single pixel.

[0028] The photovoltaic current and open-circuit voltage generation section includes: each photovoltaic conversion unit 141 contains A series of InGaAs photovoltaic diodes (typically 3-5) connected in series. Under near-infrared light illumination, the pixels generate photocurrent. Follow the formula: ,in, For the external quantum efficiency of the system, Let be the photon energy, and q be the charge carried by each charge carrier. For an InGaAs diode at a wavelength of 850 nm, >70%; The energy equivalent of each photon. The open-circuit voltage established at the pixel by this photogenerated current. Following diode characteristics: ,in, Thermoelectric voltage, This is the reverse saturation current of the diode. By selecting a low... High-quality materials and optimization This allows the pixel to generate an open-circuit voltage of 0.9-1.5V under typical operating light intensity, providing sufficient voltage to drive subsequent loads.

[0029] The electrode-tissue electrochemistry and stimulation safety section includes: the electrical signal output from the photovoltaic conversion unit is applied to brain tissue via Ptgray electrodes. The electrode-tissue interface is described using an electrochemical equivalent circuit (Randles model), including the double-layer capacitance Cdl, charge transfer resistance Rct, and tissue equivalent resistance Re. The system is designed with biphasic charge-balanced pulses as the stimulation waveform to ensure that the net injected charge is zero in each stimulation cycle, which is the basis for achieving electrochemical safety. Specific safety boundaries are controlled by the following parameters: charge density safety: the charge amount per phase pulse Q = ∫i(t)dt. This is achieved by limiting the pulse amplitude. With width To ensure the charge density Q / Ae, where Ae is the electrode surface area, is far below the safe reversible charge injection limit of the electrode material used. Thermal safety: A short pulse (e.g., 10μs) and low duty cycle operating mode is adopted. Through thermal simulation and experiments (e.g., using fiber optic temperature probes), it is verified that at the maximum designed optical power, the tissue temperature rise ΔT is controlled to be less than 1°C at a distance of 0.5-1.0 mm from the pixel surface. System monitoring: The safety monitoring module monitors the projected light power, calculates the stimulation charge density, and estimates the temperature rise in real time, ensuring that all operating points are within the above-mentioned safety window.

[0030] refer to Figure 2 This embodiment further discloses a method for visual reconstruction in total blindness, the method comprising the following steps: S100: An external image processing unit acquires environmental images and performs feature optimization processing on these images to generate an optical stimulus pattern adapted to the visual cortex's perception. Specifically, firstly, a wide-angle camera integrated into the front of the patient's glasses continuously acquires optical images of the external environment, forming a video stream. This video stream is transmitted to the visual processing unit in real time. The processing follows strict real-time constraints, with the total end-to-end latency controlled within 10 milliseconds. The processing unit performs adaptive brightness normalization and local contrast enhancement on the input raw image as basic preprocessing. Subsequently, a convolutional neural network model trained based on the YOLOv8 framework is invoked to perform edge and saliency feature detection on the preprocessed image, quickly extracting key contours and object information in the scene. This step aims to simulate and match the physiological characteristics of the primary visual cortex (V1) in processing basic shape features. To further enhance the visual details available for reconstruction, the processing unit then employs a super-resolution reconstruction algorithm based on the ESRGAN model to enhance the image resolution. After completing the aforementioned feature optimization, the software converts the processed two-dimensional image into a spatial intensity distribution matrix based on a preset mapping relationship calibrated individually by functional magnetic resonance imaging (fMRI). This mapping relationship ensures that every spatial location in the image accurately corresponds to a specific photovoltaic pixel unit in a flexible photovoltaic stimulation array implanted in the V1 cortex. The resulting intensity distribution matrix is ​​defined as the "optical stimulation pattern adapted to visual cortex perception." During the initial training phase of the system, parameters such as edge detection threshold and contrast enhancement intensity in this step can be dynamically adjusted through a closed-loop mechanism based on patient behavioral feedback to gradually optimize visual perception quality.

[0031] S200: Based on the optical stimulus pattern, a sub-threshold near-infrared light pre-tuning frame is generated, and the optical stimulus pattern is converted into a stimulus-type near-infrared light signal through a near-infrared projection unit. Specifically, firstly, based on the generated current frame optical stimulus pattern and the target pattern of the next frame, the sub-threshold near-infrared light pre-tuning frame is calculated using a pre-tuning frame generation algorithm. The core of this algorithm is to solve a regularized least-squares optimization problem, the closed-form solution of which is given by the formula x=(UᵀU+λI)⁻¹Uᵀv. Where, v is the target electric field vector determined according to the target pattern of the next frame, which is expected to be generated at a specific layer (such as layer 4 / 5) of the V1 cortex neuron; U is the photoelectric response matrix obtained in advance through finite element simulation (such as COMSOL) or experimental calibration, describing the contribution of each projected light point to the cortical electric field; λ is the regularization coefficient, used to prevent the solution from overfitting and control the pre-tuning light intensity amplitude. The calculated vector x represents the light intensity distribution of the pre-tuned frame. Its value is designed so that pixels in the flexible photovoltaic stimulation array that will be defined as dark in the next frame's stimulation pattern can receive a specific sub-threshold near-infrared illumination, generating a bias voltage of approximately 0.3V on their photovoltaic diodes, preparing for electric field focusing in subsequent stimulation frames. Subsequently, the complete optical stimulation pattern data stream, containing temporal information, is transmitted to the near-infrared projection unit. The core of this unit is an array of 1000×1000 independently addressable vertical-cavity surface-emitting lasers (VCSELs) with a center wavelength of 850nm (deviation ±10nm). The projection unit operates at a frame rate of at least 1000 frames per second (fps). Within a complete display cycle, the calculated pre-tuned frame light intensity distribution x is first converted into a sub-threshold near-infrared light pulse with a duration of 10 milliseconds and projected. After the pre-tuned frame ends, the optical stimulation pattern representing the visual information of the current frame is converted into a higher-intensity "stimulation-type" near-infrared light signal with a duration of 10 microseconds and projected. The entire spatiotemporal conversion and output of the optical signal is controlled by a dedicated drive circuit to ensure strict timing accuracy. The projection unit also integrates a micromirror alignment component based on microelectromechanical systems (MEMS). This component dynamically adjusts the beam direction based on real-time eye tracking and the fixed spatial coordinates of the cranial window, ensuring that the aforementioned near-infrared light sequence can accurately pass through the optically transparent cranial window component with an error of less than 50 micrometers and cover the corresponding photovoltaic array area under the skin.

[0032] S300: Near-infrared pre-tuned frames and stimulating near-infrared light signals are projected onto an optically transparent cranial window assembly according to a preset timing sequence. After being transmitted collaboratively by the optically transparent cranial window assembly and the scalp-adaptive optical patch, they are precisely incident on a flexible photovoltaic stimulation array. In this embodiment, the projection and transmission process of the near-infrared pre-tuned frames and stimulating near-infrared light signals follows a preset dual-frame working mechanism and precise spatiotemporal control. Specifically, operating at a 60Hz refresh rate, a pre-tuned frame is first output within each frame cycle, using 850nm near-infrared light with a sub-threshold intensity, which is precisely projected onto the central region of the optically transparent cranial window assembly through a VCSEL array and a MEMS micromirror assembly; then, the frame switches to the stimulation frame, emitting high-energy near-infrared light. The two are phase-locked, and the timing jitter is controlled within ±1μs. The optically transparent cranial window assembly consists of a three-layer structure: an outer titanium alloy fixing frame, a middle PDMS optical film, and an inner scalp-adaptive optical patch. During the preoperative preparation phase, a low-energy laser is used to thin the scalp area corresponding to the cranial window to a thickness of 0.5 mm, significantly reducing light scattering. When the near-infrared light sequence is incident, Fresnel reflection loss is controlled to within 2% after passing through the refractive index matching interface between the scalp-adaptive optical patch and the PDMS film, while the total optical transmittance remains above 90%. To ensure spatial accuracy, a MEMS micromirror alignment component is used, and real-time eye tracking and PID closed-loop control dynamically compensate for minute head movements. The light signal transmitted through the cranial window enters the subdural space and precisely illuminates the flexible photovoltaic stimulation array. Each pixel contains 3-5 InGaAs photovoltaic diodes connected in series. When receiving the pre-tuned frame light signal, the dark pixel is pre-charged to a bias voltage of 0.3V.

[0033] The aforementioned optical transmission link ensures that the activation threshold requirements of V1 neurons are met within a safe thermal load range. During the post-implantation validation phase, the optical transmission efficiency and thermal safety were verified through experiments using fiber optic temperature probes and infrared thermal imagers. This enabled efficient wireless photoelectric conversion from outside the body to cortical neurons without the need for scalp-penetrating wires, thus restoring functional vision for totally blind patients.

[0034] S400: After receiving near-infrared light signals, the flexible photovoltaic stimulation array converts the light energy into electrical stimulation signals through a photovoltaic conversion unit. These electrical stimulation signals are then applied to functional neurons in the primary visual cortex V1 via stimulation electrodes. Specifically, the photoelectric conversion and neural stimulation process of the flexible photovoltaic stimulation array is achieved through a precise opto-electric-neural coupling mechanism. When the near-infrared light stimulation sequence transmitted through the optically transparent cranial window component reaches the surface of the flexible photovoltaic stimulation array, the photovoltaic conversion unit converts the light energy into electrical stimulation signals through an N-string photovoltaic design. When receiving high-energy light signals from the stimulation frame, the photocurrent is generated according to the formula... It is confirmed that, among them, For the external quantum efficiency of the system, Let be the photon energy, and q be the charge carried by each charge carrier. For an InGaAs diode at a wavelength of 850 nm, >70%; The energy equivalent of each photon. The open-circuit voltage established at the pixel by this photogenerated current. Following diode characteristics: ,in, Thermoelectric voltage, This is the reverse saturation current of the diode. By selecting a low... High-quality materials and optimization This allows the pixels to generate an open-circuit voltage of 0.9-1.5V under typical operating light intensity, providing sufficient voltage to drive the subsequent load. This photogenerated voltage drives the current through the connected Ptgray electrode and the brain tissue load. By controlling the width and amplitude of the light pulses, the system ensures that the output electrostimulation signal is a precise biphasic charge-balanced pulse. The charge density of each pulse is strictly designed to ensure that it is below the safe reversible charge injection limit of the platinum electrode material, meeting long-term electrochemical safety requirements. This electrostimulation signal is applied directly to the V1 cortical region covered by the array through the electrode-tissue interface, mainly acting on the pyramidal neurons in layers 4 / 5 of the cortex, causing them to produce local depolarization and induce action potentials, thereby forming neural activity corresponding to the original light signal pattern. In this embodiment, the overall size of the flexible photovoltaic stimulation array is selected as 2 cm × 2 cm, containing 2000 pixel units arranged in a honeycomb pattern with a spacing of 50 micrometers, ensuring that the electrostimulation has sufficient spatial resolution to support functional visual reconstruction.

[0035] S500: Monitors visual reconstruction effectiveness through a closed-loop feedback mechanism, dynamically adjusting optical stimulation pattern parameters and electrical stimulation signal parameters until visual reconstruction achieving a preset equivalent visual acuity is achieved. Specifically, during the initial system usage and subsequent adaptive training phases, a dedicated closed-loop feedback adjustment program is initiated. This program monitors the initial effectiveness of visual reconstruction in a multimodal manner: First, it records the subject's reaction accuracy and response time as behavioral indicators. Simultaneously, it collects the patient's scalp electroencephalogram (EEG) or visual evoked potential (VEP) signals to assess the intensity and spatial distribution of the cortical neural response to specific stimulation patterns. The above feedback data is transmitted back to the external image processing unit in real time. The adaptive algorithm within the processing unit dynamically adjusts key parameters based on this feedback to optimize perceptual effects. Specifically: if the patient has difficulty recognizing a certain type of shape, the algorithm enhances the edge detection threshold or contrast of the corresponding features in image processing; if the VEP signal shows a weak response in a specific area, the algorithm fine-tunes the mapping relationship between the image pixels corresponding to that area and the photovoltaic array, or moderately increases the light intensity of the stimulation frame within safe limits on that channel. Meanwhile, the safety monitoring module ensures that all adjustments are made within a preset safety window, including stimulus charge density, projected light power, and estimated temperature rise. This closed-loop process of "stimulus-perception-feedback-adjustment" continues. The training application follows a progressive difficulty increase scheme, starting by guiding the patient to perceive a single light point and gradually transitioning to recognizing complex graphics and text. Through iterative optimization, the patient's visual perception ability eventually reaches a preset functional level, namely, achieving an equivalent visual acuity of no less than 20 / 200, thus completing the practical visual reconstruction from basic light perception to readability and navigation.

[0036] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions that cause a computer to execute the methods provided in the above-described embodiments.

[0037] Those skilled in the art will understand that all or part of the steps of the above-described method implementation can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above-described method implementation. The aforementioned storage medium includes various storage media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0038] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0039] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of each embodiment or some parts of the embodiments.

[0040] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A wireless photovoltaic visual cortex prosthesis system for total blind visual reconstruction, characterized in that, The system includes: An external image processing unit is used to acquire environmental images and generate corresponding optical stimulus patterns; A near-infrared projection unit, connected to the external image processing unit, is used to convert the optical stimulation pattern into a near-infrared light stimulation sequence. An optically transparent cranial window assembly is implanted in a predetermined location in the subject's skull and includes a light-transmitting structure optically coupled to the skull, which transmits the near-infrared light stimulation sequence to the visual cortex. A flexible photovoltaic stimulation array, implanted in the functional area of ​​the primary visual cortex V1, includes multiple photovoltaic conversion units and corresponding stimulation electrodes. The flexible photovoltaic stimulation array is used to receive the near-infrared light stimulation sequence through the optically transparent cranial window assembly and convert it into an electrical stimulation signal to stimulate the primary visual cortex V1. The near-infrared projection unit and the flexible photovoltaic stimulation array are wirelessly coupled via the optically transparent cranial window.

2. The wireless photovoltaic visual cortex prosthesis system for holistic blind vision reconstruction of claim 1, wherein, The flexible photovoltaic stimulation array is implanted at a depth of 1.0-2.0 mm below the dura mater of the visual cortex, and the flexible photovoltaic stimulation array is fixed to the curved surface of the visual cortex through a PEG-DA gel coating.

3. The wireless photovoltaic visual cortex prosthesis system for holistic blind vision reconstruction of claim 1, wherein, Each of the photovoltaic conversion units includes multiple photovoltaic diodes connected in series, and the photovoltaic diodes are made of InGaAs material.

4. The wireless photovoltaic visual cortical prosthesis system for total blindness visual reconstruction according to claim 3, characterized in that, The electrical stimulation signal output by the photovoltaic conversion unit is a biphase charge balance pulse; in the biphase charge balance pulse, the charge density of each phase pulse does not exceed the safe reversible charge injection limit of the material used for the stimulation electrode.

5. The wireless photovoltaic visual cortical prosthesis system for total blindness visual reconstruction according to claim 1, characterized in that, The optically transparent craniotomy assembly includes a titanium alloy fixation device and a PDMS optical film. The PDMS optical film covers the inner side of the titanium alloy fixation device, and the PDMS optical film has a transmittance of more than 90% for near-infrared light with a wavelength of 850-940nm. The optically transparent craniotomy assembly also includes a scalp-adaptive optical patch covering the scalp surface, wherein the refractive index of the scalp-adaptive optical patch is designed to match the refractive index of the PDMS optical film.

6. The wireless photovoltaic visual cortical prosthesis system for total blindness visual reconstruction according to claim 1, characterized in that, The external image processing unit includes an image optimization module, which runs edge detection algorithms and super-resolution reconstruction algorithms. The external image processing unit is also equipped with a pre-adjustment frame generation module, which generates a sub-threshold near-infrared light pre-adjustment frame to optically pre-charge the photovoltaic conversion unit in the flexible photovoltaic stimulation array that will darken in the next frame.

7. The wireless photovoltaic visual cortical prosthesis system for total blindness visual reconstruction according to claim 1, characterized in that, The near-infrared projection unit has a near-infrared light output wavelength of 850-915nm and a projection frame rate of not less than 1000fps. The near-infrared projection unit also includes a MEMS micromirror alignment component, which is used to accurately project the near-infrared light stimulation sequence onto the optically transparent cranial window component, with a projection alignment error of less than 50μm.

8. The wireless photovoltaic visual cortical prosthesis system for total blindness visual reconstruction according to claim 1, characterized in that, The system also includes a safety monitoring module, which is used to monitor in real time the temperature changes of the flexible photovoltaic stimulation array, the charge density of the electrical stimulation signal, and the output power of the near-infrared projection unit.

9. A method for visual reconstruction in total blindness, characterized in that, The method includes: An environmental image is acquired by an external image processing unit, and the environmental image is subjected to feature optimization processing to generate an optical stimulation pattern adapted to the visual cortex perception. Based on the optical stimulation pattern, a sub-threshold near-infrared light pre-adjustment frame is generated, and the optical stimulation pattern is converted into a stimulation-type near-infrared light signal through a near-infrared projection unit. Near-infrared light pre-frame and stimulating near-infrared light signal are projected onto the optically transparent cranial window component according to a preset time sequence. After being transmitted in coordination with the optically transparent cranial window component and the scalp-adaptive optical patch, the signal is precisely incident onto the flexible photovoltaic stimulation array. After receiving near-infrared light signals, the flexible photovoltaic stimulation array converts the light energy into electrical stimulation signals through a photovoltaic conversion unit. The electrical stimulation signals are then applied to functional neurons in the primary visual cortex V1 via stimulation electrodes. The visual reconstruction effect is monitored through a closed-loop feedback mechanism, and the parameters of the optical stimulation pattern and the electrical stimulation signal are dynamically adjusted until the visual reconstruction of the preset equivalent visual acuity is achieved.

10. The method for visual reconstruction in total blindness according to claim 9, characterized in that, The generation of sub-threshold near-infrared light pre-adjustment frames includes: Based on the mapping relationship between the optical stimulation pattern and the target cortical electric field, the optimal light intensity distribution for pre-biasing each photovoltaic conversion unit is calculated by solving a regularized least squares optimization problem.

Images