System and method for reconstructing visual cortex color vision function based on three-phase clamping waveform and terminal

Abstract

This invention provides a system, method, and terminal for visual cortex color vision reconstruction based on a three-phase clamped waveform. It is equipped with an electrode array, including an active nuclear electrode, a shielding electrode, and an isolation electrode. The system integrates a CMOS switch matrix, a residual voltage monitor, and a central processing control module. The monitor acquires the residual polarization voltage of the stimulus load in real time. The central processing module combines visual images and polarization feedback voltage to issue commands, controlling the switch matrix to output an asymmetric envelope three-phase closed-loop stimulus waveform, achieving accurate color vision reconstruction. This invention integrates AWG arbitrary waveform generation, CMOS programmable gating, and RVM residual voltage monitoring hardware to build a four-dimensional dynamic charge balance system. Relying on spatial electric field potential well constraint and microsecond-level active clamping cross-frame modulation, it constructs an electrochemically neutral tissue impedance baseline, overcoming the technical shortcomings of traditional electrostimulation's spatiotemporal separation.

Description

Technical Field

[0001] This invention relates to the field of electrophysiological stimulation technology, and in particular to a system, method and terminal for visual cortex color vision function reconstruction based on tri-phase clamping waveforms. Background Technology

[0002] In the field of visual neural prostheses and in vivo electrical stimulation of the primary visual cortex (V1), the number of high-density microelectrode arrays (MEAs) channels is growing exponentially. The industry's R&D bottleneck has shifted from early macroscopic electrode channel matching to the level of electrode-brain tissue microscopic electromagnetic field modulation and electrochemical safety control. For high-fidelity color vision reconstruction scenarios targeting functional zones of the V1 cortex, traditional monophasic and symmetrical biphasic pulse electrical stimulation paradigms have inherent adaptation defects. They cannot take into account the triple physiological constraints of the visual cortex's specialized anatomical structure, antagonistic pathway neural encoding, and electrode interface electrochemical homeostasis, making it difficult to achieve pure, safe, and controllable artificial color perception.

[0003] Currently, high-density electrical stimulation of the visual cortex is generally constrained by two fundamental physical and electrochemical challenges, becoming a prerequisite for accurate color reconstruction. On the one hand, living brain tissue is an excellent volume conductor, and conventional monopolar and bipolar stimulation currents diffuse disorderly along the interstitial spaces. The volume conduction effect significantly expands the activated tissue volume (VTA), reducing the spatial resolution of electrical stimulation and making it impossible to achieve targeted activation of the CO Blob microscopic functional area. On the other hand, the electrode-tissue interface (ETI) has strong nonlinear coupling characteristics. The Faraday transfer impedance of the interface and the dynamic coupling of the double-layer capacitance will distort the standard excitation waveform output by the back-end DAC, resulting in a mismatch between the preset electrical stimulation command and the actual electric field of the brain tissue, and a complete loss of control over the precision of color neural coding regulation. Meanwhile, traditional stimulation waveforms have two major drawbacks: they lack electric field constraint under high-density array conditions, and current can easily overflow into the Interblob contour functional region, inducing color mixing and contour perception crosstalk; they can only achieve macroscopic time-domain charge balance, and due to the influence of interface Faraday asymmetric reaction and protein adsorption capacitance hysteresis, residual polarization potential remains at the interface after stimulation. Under high-frequency stimulation, charge continues to accumulate, which can easily cause pH drift, electrode corrosion and long-term neurotoxicity.

[0004] Unlike general electrical stimulation, V1 cortical color perception reconstruction has its own strict boundary constraints, and existing stimulation architectures cannot adapt to its specific anatomy and coding rules. The primate V1 cortex is highly differentiated, with the 250 μm diameter CO Blob patch dedicated to encoding hue and saturation color information, and the surrounding interblob area dedicated to processing contour and spatial frequency information. Existing solutions lack the ability to perform closed-loop zonal mapping and electric field isolation, making it impossible to achieve 100% accurate activation of color vision targets and zero-interference isolation of contour pathways. The stepless gradation of hue and saturation relies on submillimeter-level multi-electrode current steering to achieve the centroid shift of the electric field. Traditional single-electrode fixed-point stimulation cannot replicate the neuronal cluster coding mechanism. In addition, under DKL antagonistic color space coding, increasing color saturation requires increasing local current density, which will simultaneously increase the energy of the luminance pathway. Existing paradigms cannot achieve decoupling of hue and luminance energy, and high-saturation color vision is prone to luminance flicker noise.

[0005] From the microscopic mechanism of the electrode interface, the residual polarization and waveform distortion of traditional biphase waveforms have clear physical roots. The implanted microelectrode interface follows the Randles equivalent circuit model, consisting of series tissue resistance. Interfacial double-layer capacitor Parallel charge transfer impedance The coupling structure causes the total interfacial impedance to dynamically and nonlinearly drift with stimulation frequency and charge density, directly resulting in excitation waveform distortion. Cortical electrical stimulation is divided into two safety boundaries: reversible capacitive injection and irreversible Faraday response. Capacitive stimulation relying solely on the charging and discharging of the double layer causes no tissue damage; however, once the interfacial potential exceeds the electrochemical window, irreversible damage such as hydrolysis and electrode dissolution will be triggered. In an ideal linear circuit, the biphasic pulse charge integral returns to zero. However, the interfacial polarization potential cannot be completely cleared in the in vivo environment, and the residual charge accumulates frame by frame, inducing electrochemical damage. This is the core design motivation for adding a third phase potential clamping phase and optimizing the traditional stimulation architecture. Summary of the Invention

[0006] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a system, method and terminal for visual cortical color vision function reconstruction based on triphasic clamping waveforms. This invention addresses the limitations of existing traditional biphasic cortical electrical stimulation paradigms, which are constrained by brain tissue volume conduction effects and the nonlinear electrochemical characteristics of the electrode-tissue interface. These limitations prevent the precise targeting and isolation of the V1 cortical CO Blob color vision functional area and the Interblob contour area to achieve decoupled color and brightness encoding. Furthermore, they cannot completely eliminate residual polarization charges at the interface, resulting in color vision noise pollution, insufficient color control precision, and long-term implantation electrochemical neurotoxicity defects. Consequently, these methods are unable to meet the high-fidelity, safe and pure artificial color vision reconstruction requirements under high-density arrays.

[0007] To achieve the above and other related objectives, this invention provides a visual cortex color vision function reconstruction system based on a three-phase clamping waveform. The system includes: an electrode array comprising an active core electrode, multiple shielding electrodes, and multiple isolation electrodes; wherein the active core electrode is disposed in the central region of the visual cortex color vision function target area; the multiple shielding electrodes are adjacent to the active core electrode and located within the visual cortex color vision function target area, collectively forming a shielding ring; and the multiple isolation electrodes are located in the interval regions of the visual cortex color vision function area; the physiological tissue between the active core electrode and each shielding electrode constitutes a stimulation load. The CMOS switch matrix includes: a first constant current driver connected to the active core electrode and a second constant current driver connected to the shielding electrode respectively; a residual voltage monitor connected to the active core electrode and the shielding electrode for real-time sampling of the residual polarization voltage of each stimulus load; and a central processing and control module connected to the residual voltage monitor for generating color vision control commands based on the acquired visual images and residual polarization voltages, controlling the CMOS switch matrix to execute a three-phase closed-loop stimulation waveform on the stimulus load to realize the reconstruction of color vision function in the visual cortex; wherein the three-phase closed-loop stimulation waveform is an asymmetric waveform envelope.

[0008] In one embodiment of the present invention, the three-phase closed-loop stimulation waveform corresponds to three independent stages executed sequentially, including: an active stimulation and spatial clamping stage, an active charge recovery and cross-frame adaptive modulation stage, and a constant potential clamping and zeroing stage; wherein, the waveforms of the active stimulation and spatial clamping stage and the active charge recovery and cross-frame adaptive modulation stage contain asymmetric waveform envelopes with ramp and / or step features.

[0009] In one embodiment of the present invention, the step of generating color vision control instructions based on the acquired visual images and residual polarization voltage, and controlling the CMOS switch matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load, includes: mapping the acquired visual image frames to the DKL color space and extracting DKL logic parameters; generating a multi-dimensional stimulation matrix based on the DKL logic parameters and the residual polarization voltage measured after the end of the previous frame stimulation using a waveform scheduler; decomposing the stimulation matrix into multiple time steps using arbitrary waveform generation logic to generate color vision control instructions, and controlling the CMOS switch matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load using a hardware state machine.

[0010] In one embodiment of the present invention, when generating the multidimensional stimulation matrix, the central processing and control module combines the residual polarization voltage measured after the end of the previous frame stimulation to locally correct the parameters of the active charge recovery and cross-frame adaptive modulation stages in the multidimensional stimulation matrix; wherein, the parameters include: current amplitude or duration.

[0011] In one embodiment of the present invention, the execution of a three-phase closed-loop stimulation waveform on the stimulation load by controlling the CMOS switch matrix through a hardware state machine includes: in the active stimulation and spatial clamping stage, the hardware state machine controls the CMOS switch matrix to connect the active core electrode to the constant current source cathode and the shielding electrode to the constant current source anode, forming a steep local electric field potential well in the cytochrome oxidase plaque region, while the arbitrary waveform generator outputs an asymmetric waveform with programmable rising edges and ramp and step characteristics; in the active charge recovery and cross-frame adaptive modulation stage, the hardware state machine controls the CMOS switch matrix to connect the active core electrode to the constant current source anode and the shielding electrode to the constant current source cathode; in the constant potential clamping and clearing stage, the hardware state machine controls the CMOS switch matrix to cut off the constant current drive circuit.

[0012] In one embodiment of the present invention, the central processing and control module controls the CMOS switch matrix to drive a designated active core electrode and a shielding electrode for stimulation through a constructed routing table; wherein, the hardware routing table includes: electrode ID, physical coordinates, mask label, associated active electrode ID, number of shielding ring electrodes, switch status code, and abnormal flag bit.

[0013] In one embodiment of the present invention, the method of constructing the routing table includes: sending biphase constant current probe pulses to the electrode array channel by channel in a time-division multiplexing manner through a CMOS switch matrix, and performing residual voltage active clamping after the pulse to collect multi-dimensional physiological characteristics of each channel; processing the multi-dimensional physiological characteristics and generating a binary mask that can distinguish the interval region between cytochrome oxidase plaques and adjacent cytochrome oxidase plaques by combining an adaptive dual threshold algorithm; performing initial screening of all channels of electrodes based on the binary mask, isolating non-functional area electrodes, and splitting multiple independent CO Blob functional connected domains through the 8-neighborhood connectivity rule; calculating the number of shieldable electrodes in each CO Blob functional connected domain, and selecting active core electrodes and shielding electrodes accordingly, marking non-connected domain electrodes as isolation electrodes and setting them to a high-resistance state; and constructing a hardware routing table based on the active core electrodes, shielding electrodes, and isolation electrodes.

[0014] In one embodiment of the present invention, the step of calculating the number of shieldable electrodes within each CO Blob functional connectivity region and selecting the active core electrode and shielding electrode accordingly includes: calculating the number of shielding resources for each electrode within each connectivity region; wherein the number of shielding resources is defined as the number of other electrodes belonging to the same connectivity region in the neighborhood of the electrode with a preset radius as the center; traversing the number of shielding resources of all electrodes in each connectivity region and selecting the electrode with the largest number of shielding resources as the active core electrode of the connectivity region; marking all electrodes belonging to the same connectivity region in the neighborhood of the active core electrode as shielding electrodes, thereby forming a shielding ring.

[0015] To achieve the above and other related objectives, this invention provides a method for visual cortex color vision function reconstruction based on a three-phase clamping waveform, applied to the aforementioned visual cortex color vision function reconstruction system based on a three-phase clamping waveform. The method includes: acquiring a collected visual image and a residual polarization voltage detected by a residual voltage monitor; generating a color vision control command based on the acquired visual image and the residual polarization voltage detected by the residual voltage monitor, controlling a CMOS switching matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load, thereby realizing visual cortex color vision function reconstruction; wherein, the three-phase closed-loop stimulation waveform is an asymmetric waveform envelope.

[0016] To achieve the above and other related objectives, the present invention provides an electronic terminal, comprising: one or more memories and one or more processors; the one or more memories being used to store a computer program; and the one or more processors being connected to the memories and used to run the computer program to execute the method.

[0017] As described above, this invention is a visual cortex color vision function reconstruction system, method, and terminal based on a three-phase clamped waveform, which has the following beneficial effects: This invention incorporates an electrode array, which includes an active core electrode, a shielding electrode, and an isolation electrode. The active core electrode is positioned in the central region of the visual cortex color vision function target area. The shielding electrode is positioned adjacent to the active core electrode within the visual cortex color vision function target area, forming an electric field shielding ring. The isolation electrode is positioned in the interval region of the visual cortex color vision function area. The physiological tissue between the active core electrode and the shielding electrode constitutes the stimulation load. The system integrates a CMOS switch matrix, a residual voltage monitor, and a central processing control module. The switch matrix drives the active core electrode and the shielding electrode respectively through independent constant current drivers. The residual voltage monitor collects the residual polarization voltage of the stimulation load in real time. The central processing control module combines visual image information and polarization feedback voltage to issue commands, controlling the switch matrix to output an asymmetric envelope three-phase closed-loop stimulation waveform, thus completing accurate color vision reconstruction. This invention integrates AWG arbitrary waveform generation, CMOS programmable gating, and RVM residual voltage monitoring hardware to build a four-dimensional dynamic charge balance system. Relying on spatial electric field potential well constraint and microsecond-level active clamping cross-frame modulation, it constructs an electrochemical neutral tissue impedance baseline, breaking through the technical shortcomings of traditional electrostimulation with its spatiotemporal separation. Attached Figure Description

[0018] Figure 1 The diagram shown is a structural schematic of a visual cortex color vision function reconstruction system based on a three-phase clamping waveform according to an embodiment of the present invention.

[0019] Figure 2 The diagram shown is a schematic representation of the electrode array arrangement in one embodiment of the present invention.

[0020] Figure 3The diagram shown is a circuit schematic of a constant current driver according to an embodiment of the present invention.

[0021] Figure 4 The diagram shown is a schematic representation of a CO Blob connected region in one embodiment of the present invention.

[0022] Figure 5 The diagram shown is a schematic representation of a three-phase closed-loop stimulation waveform in one embodiment of the present invention.

[0023] Figure 6 The diagram shown is a flowchart illustrating a method for reconstructing color vision function of the visual cortex based on a three-phase clamping waveform according to an embodiment of the present invention.

[0024] Figure 7 The diagram shown is a structural schematic of an electronic terminal according to an embodiment of the present invention. Detailed Implementation

[0025] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0026] It should be noted that in the following description, reference is made to the accompanying drawings, which illustrate several embodiments of the invention. It should be understood that other embodiments may also be used, and changes in mechanical composition, structure, electrical system, and operation may be made without departing from the spirit and scope of the invention. The following detailed description should not be considered limiting, and the scope of the embodiments of the invention is defined only by the claims of the published patents. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. Spatially related terms, such as “upper,” “lower,” “left,” “right,” “below,” “below,” “lower part,” “above,” “upper part,” etc., may be used herein to illustrate the relationship between one element or feature shown in the figures and another element or feature.

[0027] Throughout this specification, when it is said that a part is "connected" to another part, this includes not only "direct connection" but also "indirect connection" by placing other elements in between. Furthermore, when it is said that a part "includes" a certain constituent element, unless otherwise stated otherwise, this does not exclude other constituent elements, but rather means that other constituent elements may also be included.

[0028] The terms "first," "second," and "third," etc., used herein are for the purpose of describing various parts, components, regions, layers, and / or segments, but are not limiting. These terms are used only to distinguish one part, component, region, layer, or segment from others. Therefore, the "first part," "component," "region," "layer," or "segment" described below may refer to a "second part," "component," "region," "layer," or "segment" without departing from the scope of this invention.

[0029] Furthermore, as used herein, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context indicates otherwise. It should be further understood that the terms “comprising,” “including,” indicate the presence of the stated feature, operation, element, component, item, kind, and / or group, but do not preclude the presence, occurrence, or addition of one or more other features, operations, elements, components, items, kinds, and / or groups. The terms “or” and “and / or” as used herein are interpreted as inclusive, or mean any one or any combination thereof. Thus, “A, B, or C” or “A, B, and / or C” means “any one of: A; B; C; A and B; A and C; B and C; A, B, and C.” Exceptions to this definition arise only when combinations of elements, functions, or operations are inherently mutually exclusive in some manner.

[0030] This invention provides a visual cortex color vision reconstruction system based on a three-phase clamped waveform. The system incorporates an electrode array comprising an active core electrode, a shielding electrode, and an isolation electrode. The active core electrode is positioned at the center of the target area for color vision in the visual cortex. The shielding electrode is positioned adjacent to the active core electrode within the target area, forming an electric field shielding ring. The isolation electrode is positioned in the intervals within the target area. Physiological tissue between the active core electrode and the shielding electrode constitutes the stimulation load. The system integrates a CMOS switch matrix, a residual voltage monitor, and a central processing control module. The switch matrix drives the active core electrode and the shielding electrode via independent constant current drivers. The residual voltage monitor collects the residual polarization voltage of the stimulation load in real time. The central processing control module combines visual image information and polarization feedback voltage to issue commands, controlling the switch matrix to output an asymmetric envelope three-phase closed-loop stimulation waveform, thus achieving accurate color vision reconstruction. This invention integrates AWG arbitrary waveform generation, CMOS programmable gating, and RVM residual voltage monitoring hardware to build a four-dimensional dynamic charge balance system. Relying on spatial electric field potential well constraint and microsecond-level active clamping cross-frame modulation, it constructs an electrochemical neutral tissue impedance baseline, breaking through the technical shortcomings of traditional electrostimulation with its spatiotemporal separation.

[0031] The present invention will now be described in detail with reference to the accompanying drawings, so that those skilled in the art can readily implement it. The present invention can be embodied in many different forms and is not limited to the embodiments described herein.

[0032] like Figure 1 This invention presents a schematic diagram of a visual cortex color vision function reconstruction system based on a three-phase clamping waveform, according to an embodiment of the present invention.

[0033] The visual cortex color vision function reconstruction system based on three-phase clamping waveforms includes:

[0034] Electrode array 1 includes an active nuclear electrode, multiple shielding electrodes, and multiple isolation electrodes; wherein, the active nuclear electrode is disposed in the central region of the color vision target area of ​​the visual cortex; the multiple shielding electrodes are adjacent to the active nuclear electrode and located in the color vision target area of ​​the visual cortex, together forming a shielding ring; the multiple isolation electrodes are located in the interval regions of the color vision target area of ​​the visual cortex; the physiological tissue between the active nuclear electrode and each shielding electrode constitutes the stimulation load;

[0035] Specifically, electrode array 1 is implanted in the visual cortex, directly acting on the brain's visual cortex. The target area for color vision function in the visual cortex is the cytochrome oxidase plaque. Three types of electrodes are strategically deployed based on the spatial distribution of the cytochrome oxidase plaque (the core neural functional area of ​​the visual cortex responsible for color perception), achieving zoned signal output, electric field shielding, and regional isolation. Simultaneously, it utilizes neurophysiological tissue to form a stimulation circuit load. This includes:

[0036] Active nuclear electrodes, individually deployed and fixed at the geometric center of a single cytochrome oxidase plaque.

[0037] The shielding electrodes, arranged in multiple combinations, are located adjacent to the active nuclear electrode and all within the same cytochrome oxidase plaque, forming a closed shielding ring around the active nuclear electrode. Their function is to constrain the range of the stimulation electric field, prevent the main stimulation signal from diffusing outwards from the plaque, avoid accidental activation of adjacent neural regions, and ensure the target specificity of color vision stimulation.

[0038] The isolation electrodes, multiple of which are distributed in the blank space between adjacent cytochrome oxidase plaques, physically and electrically isolate different functional plaques, block crosstalk of electrical signals between plaques, and ensure that each color vision functional area works independently.

[0039] like Figure 2 For example, the red dashed circle represents the boundary of the cytochrome oxidase plaque (CO Blob). After the electrode array 1 is implanted, it contacts the cerebral cortex. Through the setting of the CMOS switch array and electrodes, the active nuclear electrode is deployed to the center of the CO Blob plaque, and the shielding electrode is deployed adjacent to the active nuclear electrode inside the CO Blob plaque, forming an electric field shielding ring. The isolation electrode is deployed in the interblob interval region. The active nuclear electrode is the central electrode of the red dashed circle, and the shielding electrode is also located in the red dashed circle. The blue dashed box represents the interblob region.

[0040] The living neurophysiological tissue between the active nuclear electrode and each set of shielding electrodes constitutes a dedicated stimulation load for the three-phase clamping stimulation waveform, which is the direct carrier of the three-phase electrical stimulation signal. The entire set of electrical stimulation signals is applied in a closed loop within this load circuit.

[0041] CMOS switch matrix enables path selection and precise current output, including:

[0042] The first constant current driver is dedicated to the active nuclear electrode, providing a constant excitation current to the main stimulation circuit and carrying the main output component of the three-phase waveform.

[0043] Second constant current driver: The number corresponds one-to-one with the shielding electrode. Each shielding electrode is matched with a second constant current driver to independently control the current output at each point of the shielding ring, thereby achieving balanced control of the shielding electric field.

[0044] The system relies on a CMOS switch array to achieve high-speed switching of multiple channels and adaptive loop impedance, ensuring that multiple electrodes can output steady-state current synchronously or in a time-sharing manner, supporting the orderly loading of three-phase clamped closed-loop waveforms and precise application to neural loads.

[0045] In this embodiment, the CMOS switch matrix 2 replaces the traditional simple discrete switch array, which can dynamically couple one or more sets of constant current drivers to any specified microelectrode point. It can not only flexibly build an "active core-shielded ring" topology stimulation structure, but also complete the absolute value redundancy verification and autonomous calibration of charge injection from the hardware level through the driver pairing coupling architecture, thereby improving the safety and accuracy of electrical stimulation.

[0046] like Figure 3 As shown, a first constant current driver includes a current source, a current sink, a first switch S1, a second switch S2, a third switch S3, and a fourth switch S4. The first switch S1 is connected in parallel with the series circuit of the current source and the third switch S3. The second switch S2 is connected in parallel with the series circuit of the fourth switch S4 and the current sink. One end of the first switch S1 is connected to a power source, and the other end is connected to the electrode. One end of the current source is connected to a power source, and the other end is connected to the third switch S3. One end of the third switch S3 is connected to the current source, and the other end is connected to the electrode. One end of the second switch S2 is connected to the electrode, and the other end is grounded. One end of the fourth switch S4 is connected to the electrode, and the other end is connected to the current sink. One end of the current sink is connected to the fourth switch S4, and the other end is grounded. Using a pairwise pairing method, one end of the load is connected to the electrode of the first constant current driver, and the other end of the load is connected to the electrode of the second constant current driver.

[0047] The current trap of the constant current driver is an amplifier-based current trap. This current trap may include a first amplifier OTA1, a first field-effect transistor MC1, a first current-mode digital-to-analog converter (DAC), and a second current-mode DAC. The first current-mode DAC provides the stimulation current, and the second current-mode DAC compensates for the stimulation current. The positive input terminal of the first amplifier OTA1 is a first preset voltage VB1. The negative input terminal of the first amplifier OTA1 and the source of the first field-effect transistor MC1 are both connected to the first and second current-mode DACs. The output of the first amplifier OTA1 is connected to the gate of the first field-effect transistor MC1, and the drain of the first field-effect transistor MC1 is connected to the fourth switch S4 of the constant current driver. The first field-effect transistor MC1 can be an NMOS transistor (N-Metal-Oxide-Semiconductor). The current source of the constant current driver includes a second amplifier OTA2, a second field-effect transistor MC2, a third field-effect transistor M3, a capacitor C1, and a fifth switch S5. In this circuit, one end of capacitor C1 is connected to the power supply of the constant current driver, and the other end of capacitor C1 is connected to one end of the fifth switch S5. The other end of the fifth switch S5 is connected to the drain of the second field-effect transistor MC2. The source of the third field-effect transistor M3 is connected to the power supply, and the gate of the third field-effect transistor M3 is connected between capacitor C1 and the fifth switch S5. The drain of the third field-effect transistor M3 is connected to the source of the second field-effect transistor MC2 and the negative input terminal of the second amplifier OTA2. The positive input terminal of the second amplifier OTA2 is the second preset voltage, and the output of the second amplifier (OTA2) is connected to the gate of the second field-effect transistor MC2. The drain of the second field-effect transistor MC2 is connected to the third switch S3. The second field-effect transistor MC2 and the third field-effect transistor M3 are PMOS transistors (Positive Channel Metal Oxide Semiconductors, n-type substrate, p-channel, MOS transistors that transport current by the flow of holes). Capacitor C1 is a metal capacitor with a capacitance value of, for example, 6pF. The constant current driver also includes a sixth switch S6, which is connected between the drain of the first field-effect transistor MC1 and the drain of the second field-effect transistor MC2. When electrically stimulating the load, the two constant current drivers are used in pairs; these are referred to here as a constant current driver and another constant current driver. One end of the load is connected to the first constant current driver, and the other end is connected to the second constant current driver. The constant current driver also includes a seventh switch S7, one end of which is connected to an electrode, and the other end is connected to a third preset voltage. The third preset voltage can be set to 0 volts.

[0048] The residual voltage monitor 3, connected to the active nuclear electrode and the shielding electrode, can perform real-time continuous sampling, signal-to-digital conversion, and data transmission of the residual polarization voltage at both ends of the stimulation load. Under long-term in vivo electrical stimulation, charge accumulation polarization and residual voltage accumulation can occur at the electrode-nerve tissue interface, which can easily induce irreversible nerve damage and distortion of the three-phase stimulation waveform. This module provides raw feedback data to the back-end closed-loop control unit by monitoring the residual polarization voltage in real time across the entire domain. It is the core feedback unit for the system to achieve precise three-phase closed-loop electrical stimulation.

[0049] The central processing and control module 4 is connected to the residual voltage monitor 3. It is used to generate color vision control commands based on the acquired visual images and residual polarization voltage, and control the CMOS switch matrix 2 to execute a three-phase closed-loop stimulation waveform on the stimulation load to realize the reconstruction of color vision function in the visual cortex. The three-phase closed-loop stimulation waveform is an asymmetric waveform envelope.

[0050] In one embodiment, unlike traditional implantation systems that require physical alignment of the visual cortex's microscopic anatomical structures, this system, after implanting the high-density microelectrode array 1, does not require physical alignment of artificial anatomical structures. During the initial startup phase, relying on extremely weak probing pulses across the entire array, all electrode channels are scanned, and local field potentials (LFPs) are simultaneously acquired to create an in-situ functional map of the V1 visual cortex. Combined with electrophysiological characteristics, functional regions are autonomously divided, and initial electrode labeling is completed: electrodes whose physical coordinates fall within the cytochrome oxidase plaque (CO Blob) are initially labeled as active electrodes; electrodes whose physical coordinates are adjacent to active electrodes and also located within the CO Blob are initially labeled as shielding electrodes; electrodes falling into the interblob are directly labeled as isolation electrodes and physically disconnected using hardware. After completing the full-domain evoked closed-loop mapping, the off-chip storage unit generates a low-level hardware routing table, archiving all electrode functional labeling information to provide a routing basis for subsequent precise electrical stimulation delivery. For scenarios with multiple alternative electrodes within a single CO Blob, all electrodes are uniformly marked as backup active electrodes. By counting the number of shielded electrodes that can be matched with a single backup active electrode, the electrode corresponding to the maximum number of shielded electrodes is selected as the final working active electrode.

[0051] The specific process includes:

[0052] The CMOS switch matrix 2 sends bi-phase constant current detection pulses to the electrode array 1 channel by channel in a time-division multiplexing manner, and performs residual voltage active clamping after the pulse to collect multi-dimensional physiological characteristics of each channel.

[0053] Specifically, a time-division multiplexing (TDM) mode is used to scan all electrode channels row by row and column by column. To improve scanning efficiency, parallel synchronous detection of spatially non-adjacent electrode groups is supported. A biphase constant-current probe pulse is emitted to the electrode channel, and residual voltage active clamping is performed after the pulse ends to eliminate electrode-tissue interface polarization effects and ensure signal acquisition accuracy. Local field potentials (LFPs) are acquired synchronously at a sampling rate of no less than 2kHz. The acquired signal includes pulse-induced dynamic potentials and 1-second resting-state potentials, which serve as the raw data for subsequent feature calculations.

[0054] Based on the collected electrical and electrophysiological signals, a three-dimensional feature vector f_i = (Z, A, La) is constructed as the basis for distinguishing CO plaques from interblobs.

[0055] Impedance mode characteristic Z: The electrode voltage is acquired at the midpoint of the anode phase of the probe pulse, and the 1kHz impedance mode value at the interface between the electrode and biological tissue is calculated by combining it with the known injection current. CO plaques have larger extracellular spaces and richer vascular distribution, and their overall impedance is lower than that of the plaque interstitial region; by utilizing the spatial gradient change of impedance, the physical boundary of CO plaques can be accurately defined.

[0056] Evoked LFP peak-to-peak amplitude A: A signal latency window of a preset size is extracted, and the peak-to-peak value of evoked potentials P1-N1 is calculated. The excitatory connectivity of neural circuits within CO plaques is stronger, and the thalamic neural input is denser, resulting in evoked potential amplitudes that are significantly higher than those in the septal region. This is a core feature distinguishing the two types of regions.

[0057] Logarithmic slope of power spectral density La in the δ-α band: Based on the 1-second resting-state LFP signal, the logarithmic slope of the power spectral density (PSD) in the 1~12Hz (δ-α band) is calculated to characterize the local neural rhythm activity, supplement the discrimination dimension, and improve the anti-interference ability of the partitioning algorithm.

[0058] Based on the multidimensional feature vector, a binary mask is generated by combining the adaptive dual threshold algorithm to distinguish the interval region between cytochrome oxidase plaques and adjacent cytochrome oxidase plaques.

[0059] Specifically, the spatial distribution data of the three-dimensional features is smoothed to filter out single-channel random measurement noise and restore the continuous spatial contour of the visual cortex. An adaptive double-threshold algorithm with Otsu's method and adjacent pixel continuity constraints is used to generate a binary mask M(x,y); where 1 indicates that the electrode physical coordinates are located within the Blob, and 0 indicates that they are located in the Interblob.

[0060] A preliminary screening of all-channel electrodes is performed using a binary mask, isolating non-functional area electrodes. Simultaneously, multiple independent CO Blob functional connectivity domains are extracted using an 8-neighborhood connectivity rule. Specifically, the electrodes are initially classified based on the binary mask results: a mask value of 1 indicates inclusion in the CO patch candidate region, serving as a candidate channel for active or shielding electrodes; a mask value of 0 indicates identification as an isolation electrode, and the CMOS switch matrix is ​​controlled to set this channel to a high-resistivity state. The 8-neighborhood connectivity rule is then used to locate all spatially connected Blob electrode sets. Each set corresponds to a distinguishable CO Blob region.

[0061] The number of shieldable electrodes is calculated within each CO Blob functional connectivity region, and active core electrodes and shielding electrodes are selected accordingly. Electrodes in non-connectivity regions are marked as isolation electrodes and set to a high-resistance state.

[0062] Specifically, within a single CO2 patch connectivity domain, the number of shieldable electrodes S(e) is used as the core evaluation index to classify active nuclear electrodes, shielded electrodes, and backup electrodes, and to construct an internal shielding ring. For any electrode e within the connectivity domain, a neighborhood set N(e) centered on it with a radius not exceeding a set value is defined; the number of electrodes belonging to the current connectivity domain within the set (excluding electrode e itself) is the number of shieldable electrodes S(e) for that electrode, representing the available resources for constructing a shielding ring around the electrode.

[0063] For each connected region, the electrode corresponding to the maximum value of S(e) is selected as the active core electrode; if multiple electrodes have the same S(e) value, the electrode closest to the geometric centroid of the connected region is selected to ensure that the stimulation site falls in the functional core region of CO plaque and improve the uniformity of stimulation.

[0064] Electrodes within the vicinity of the active nuclear electrode and belonging to the same connected region are uniformly marked as shielding electrodes, forming a shielding ring. The shielding ring is entirely confined within the CO plaque to prevent the electric field from leaking into the plaque interstitial region and avoid interfering with other neural pathways.

[0065] The remaining electrodes within the connected domain that were not selected as active electrodes or shielding electrodes are marked as backup active electrodes; they are sorted in descending order of S(e) value and stored in a priority table for redundant replacement in case of main electrode failure or parameter switching.

[0066] When the maximum S(e) of electrodes within a connected region is less than a preset value, it is determined that the number of electrodes inside the patch is insufficient to construct a complete and continuous shielding ring, and the following three strategies are executed according to priority:

[0067] Strategy A: Expand the shielding neighborhood;

[0068] Expand the shielding neighborhood radius, but still strictly limit electrode selection to within the current CO2 patch; if the shielding ring deployment requirements still cannot be met after expansion, switch to strategy B.

[0069] Strategy B: Activate patch spacer auxiliary shielding;

[0070] Two to four plaque-spacer electrodes within the direct neighborhood of active electrode 8 are selected as temporary shielding electrodes. These electrodes retain their original isolation properties and are only temporarily connected to the anode bus by the switching matrix during the stimulation phase; after a single stimulation phase ends, they immediately return to a high-resistivity state.

[0071] Strategy C: Determine it as an invalid region and discard it;

[0072] If the maximum S(e) in the connected region is less than the minimum value, the region is determined to be a measurement artifact or a very small CO plaque fragment, and is marked as an isolation region, no longer participating in electrical stimulation.

[0073] A hardware routing table is constructed based on the electrode classification results. After the mapping is completed, the off-chip FPGA control unit constructs the underlying hardware routing table in memory, with each record corresponding to an electrode channel.

[0074] The central processing and control module 4, relying on the underlying hardware routing table generated by the previous mapping, manages the CMOS switch matrix 2 in a targeted manner, driving the designated active core electrode and its matching shielded electrode to complete the targeted electrical stimulation operation. This module relies on the hardware routing table to realize the automated management, targeted selection, and status monitoring of the electrode channels. It is the core index for the main control module to drive the CMOS switch matrix. The functions of each field are as follows:

[0075] Electrode ID: A unique channel number used to address the corresponding CMOS switch and electrode path;

[0076] Physical coordinates: Spatial position of electrode array 1, matching the actual distribution of CO2 patches in the visual cortex, supporting the calculation of neighborhood, distance, and connected components;

[0077] Mask label: Associated with binary mask results, distinguishing whether the electrode belongs to the CO patch or the interblob region;

[0078] Associated active electrode ID: The shielded electrode is bound to the corresponding active core electrode, clearly defining its belonging to the shielding ring;

[0079] Number of shielded ring electrodes: Records the total number of shielded electrodes associated with a single active electrode, used for loop integrity verification;

[0080] Switch status coding: Defines the channel operating mode (stimulus output, sampling, high blocking on, temporary multiplexing, etc.), directly controlling the on / off state of the CMOS switch matrix;

[0081] Abnormal flag: Records calibration / operational abnormalities, such as shielding ring radius expansion, temporary shielding using interval zone electrodes, invalid patches, etc., for adaptive correction of closed-loop parameters.

[0082] In one embodiment, the three-phase closed-loop stimulation waveform corresponds to three independent stages executed sequentially, including: waveforms of the active stimulation and spatial clamping stage, the active charge recovery and cross-frame adaptive modulation stage, and the constant potential clamping and zeroing stage; wherein, the waveforms of the active stimulation and spatial clamping stage and the active charge recovery and cross-frame adaptive modulation stage contain asymmetric waveform envelopes with ramp and / or step features.

[0083] In one specific embodiment, the step of generating color vision control commands based on the acquired visual images and residual polarization voltage, and controlling the CMOS switch matrix 2 to execute a three-phase closed-loop stimulation waveform on the stimulation load, includes:

[0084] The acquired visual image frames are mapped to the DKL color space, and the DKL logic parameters are extracted. Specifically, the current visual image frame acquired by the camera is preprocessed sequentially with white balance, depixelation, and gamma correction to output a standardized sRGB image and obtain the RGB three-channel pixel values ​​of the image. Then, the sRGB image is converted to the Derrington-Krauskopf-Lennie (DKL) antagonistic color space, which is consistent with the biological color vision mechanism of primate retinal ganglion cells and lateral geniculate nucleus (LGN). The coordinate axis is built based on two native color antagonistic neural pathways, namely the LM red-green antagonistic pathway and the S-(L+M) blue-yellow antagonistic pathway, which is adapted to the brain's native color vision encoding logic. The RGB to DKL space conversion requires step-by-step calculation via the LMS cone intermediate space. The core conversion steps are: First, based on the sensor calibration of the photosensitivity matrix, the RGB pixel values ​​are converted to the cone quantum capture rate (L, M, S); Second, the LMS parameters are converted to the DKL three-axis coordinates (τ, δ, ε). The three axes are defined as follows: τ is the luminance axis, corresponding to L+M composite luminance information, transmitted through the macrocellular pathway; δ is the red-green antagonistic axis, where a positive value activates red color perception and a negative value activates green color perception; ε is the blue-yellow antagonistic axis, where a positive value activates blue color perception and a negative value activates yellow color perception. The final standardized solution outputs a triplet of DKL logic parameters: {LUM, r, θ}. The parameters are defined as follows: LUM is the luminance axis coordinate, representing the brightness and darkness of the image, not the chrominance information; r is the spatial radial distance, representing the chrominance modulation amplitude of the color relative to the origin of the equal-energy white point, corresponding to the saturation of the color stimulus; θ is the plane elevation angle, representing the angle of color offset relative to the luminance axis in the equal-luminance plane, with θ=0° corresponding to the standard equal-luminance plane.

[0085] The waveform scheduler utilizes a spatiotemporal tensor synthesis algorithm to generate a multidimensional stimulation matrix based on the DKL logic parameters and the residual polarization voltage measured after the end of the previous frame stimulus. Specifically, the central processing and control module 4 has a built-in waveform scheduler that synchronously receives the triplet DKL logic parameters and the corresponding LM antagonistic channel weight coefficient w. It also links the residual polarization voltage obtained from the closed-loop sampling of the previous frame stimulus and iteratively solves the problem using the spatiotemporal tensor synthesis algorithm to generate a multidimensional stimulation matrix I(t) containing full-time parameters of the three-phase waveform. The matrix embeds all controllable parameters such as current amplitude, duration, electric field shielding ratio, and potential clamping threshold for each stage of the three-phase waveform. The matrix parameters are corrected in a closed loop based on the residual voltage to avoid waveform distortion caused by interface polarization.

[0086] The stimulation matrix is ​​decomposed into multiple time steps using arbitrary waveform generation logic to generate color vision control instructions. A hardware state machine then controls the CMOS switch matrix 2 to execute a three-phase closed-loop stimulation waveform on the stimulation load. Specifically, the multi-dimensional stimulation matrix is ​​broken down and compiled into microsecond-level time steps using arbitrary waveform generation logic to generate standardized color vision control instructions. These instructions are parsed by the on-chip hardware state machine, linking with the underlying hardware routing table to precisely control the configuration switching of the CMOS switch matrix 2. This schedules designated active core electrodes and their corresponding shielding electrodes to work collaboratively, sequentially executing a complete three-phase closed-loop stimulation waveform on the neural stimulation load. After the system enters steady-state visual stimulation, subsequent new visual images only require iterative updates to the DKL triplet parameters and the channel weight W matrix. The waveform scheduler can directly retrieve the fixed routing table to quickly drive the corresponding functional electrodes to complete the color vision electrical stimulation output, reducing iterative computational overhead and improving the color vision reconstruction response rate.

[0087] In one embodiment, unlike the open-loop waveform output architecture, the central processing and control module 4 has cross-frame charge adaptive correction capability: During the generation of the multidimensional stimulus matrix, the central processing and control module 4 retrieves the residual polarization voltage data collected by the residual voltage monitor 3 (RVM) after the end of the previous stimulus frame, and performs local closed-loop correction on the parameters of the active charge recovery and cross-frame adaptive modulation stages within the multidimensional stimulus matrix. The correctable parameters include two core control quantities: current amplitude and pulse duration. The specific adaptive control logic is as follows: If the RVM detects a trace of negative polarization residue in the stimulus load, the built-in arbitrary waveform generator (AWG) dynamically increases the anode current pulse amplitude of the active charge recovery and cross-frame adaptive modulation stages of the current frame. Relying on the cross-frame adaptive balancing mechanism, the residual charge at the interface is neutralized in advance, avoiding the problems of multi-frame superposition charge accumulation, electrode polarization damage, and waveform shift from the source. The parameters of the other two waveform stages of the matrix remain unchanged, ensuring the stability of color vision stimulus encoding. The matrix embeds all controllable parameters of the three-phase waveform at each stage, including current amplitude, duration, electric field shielding ratio, and potential clamping threshold. It relies on residual voltage to complete closed-loop correction of matrix parameters, avoiding waveform distortion caused by interface polarization.

[0088] In one embodiment, the step of controlling the CMOS switch matrix 2 to execute a three-phase closed-loop stimulation waveform on the stimulation load via a hardware state machine includes:

[0089] During the active stimulation and spatial clamping phase, the hardware state machine controls the CMOS switch matrix 2 to connect the active nuclear electrode to the constant current source cathode and the shielded electrode constant current source to the anode, forming a steep local electric field potential well within the cytochrome oxidase plaque region; specifically, combined with Figure 2 Circuit timing details: In the pre-start state, the enable terminals of the first amplifier OTA1 and the second amplifier OTA2 are simultaneously turned on, and the fifth switch S5 and the sixth switch S6 are closed and conducting; the first field-effect transistor MC1 outputs a first current equal to the stimulation current given by the first current-type digital-to-analog converter. This current is copied to the gate of the third field-effect transistor M3 and stored in the compensation capacitor C1. After the formal start of the stage, reverse cathode electrical stimulation is performed: the active nuclear electrode corresponding to the constant current driver OTA1 is kept on, the fourth switch S4 is closed, and the shielding electrode corresponding to the constant current driver first switch S1 is closed, the circuit is connected to form a reverse stimulation current, and the amplitude of the reverse current is exactly the same as the first current output by MC1. Meanwhile, the AWG output features a programmable rising edge and an asymmetric waveform with both ramp and step characteristics, abandoning the instantaneous jump square wave format. The rising edge parameters can be controlled and configured, with a typical rising edge duration of 10μs. The working mechanism is that the active nuclear electrode outputs the DKL command to match the stimulation current, and the outer shielding ring electrode synchronously absorbs an equal amount of loop current, forming a steep local electric field potential well. The pulse width in this stage is independently programmable from 0 to 350μs.

[0090] During the active charge recovery and cross-frame adaptive modulation phase, the hardware state machine controls the CMOS switch matrix 2 to connect the active core electrode to the anode of the constant current source and the shield electrode to the cathode. Simultaneously, an arbitrary waveform generator outputs an asymmetric waveform with programmable rising edges and ramp and step characteristics. Specifically, when the load is positively stimulated (i.e., anodic stimulation), the enable terminal of the second amplifier (OTA2) of the constant current driver for the active core electrode is turned on, the third switch S3 is closed, and the second switch S2 of the constant current driver for the shield electrode is closed, thus forming a positive stimulation current. This positive stimulation current is a second current generated by the third field-effect transistor M3. The second current generated by the third field-effect transistor M3 is copied from the first field-effect transistor (MC1) and stored in the capacitor during the first mode; therefore, the second current is equal to the first current, and consequently, the positive stimulation current is equal to the reverse stimulation current. When the load is electrically stimulated, the first current generated by the first field-effect transistor (MC1) is copied to the gate of the third field-effect transistor M3 and stored in capacitor C1 by the constant current driver during the active stimulation and spatial clamping phase. This ensures that the forward stimulation current during the active charge recovery and cross-frame adaptive modulation phase is completely consistent with the reverse stimulation current during the second mode, thereby achieving excellent charge balance and avoiding the charge imbalance caused by current deviations due to different circuits during forward and reverse stimulation in existing technologies. Simultaneously, the arbitrary waveform generator outputs an asymmetric waveform with a programmable rising edge ramp.

[0091] During the constant potential clamping and clearing phase, the hardware state machine controls CMOS switch matrix 2 to cut off the constant current drive circuit. Specifically, the seventh switch S7 of the constant current driver for the active core electrode and the constant current driver for the shield electrode closes simultaneously to release excess charge on the load. This further improves the charge balance of the constant current driver. The system does not passively wait for the double-layer capacitance (…). Instead of allowing a prolonged natural discharge, the circuit is driven to a specific constant reference potential, such as 0. This reference potential is monitored in real time by the RVM, and processing of the next frame only proceeds when the voltage zeroing requirement is met. The RVM performs high-precision sampling of polarization residues and updates the registers for use in the next frame.

[0092] To better describe the visual cortex color vision function reconstruction system, the following specific embodiments will be used for illustration.

[0093] Example: Red light perception reconstruction

[0094] In this embodiment, a red apple image was selected as the visual input sample. The CIE 1931 standard chromaticity coordinates of this sample are x=0.64 and y=0.33, corresponding to a dominant wavelength of approximately 630nm. The experiment used an invasive electrode array implanted in the primary visual cortex (V1 region) of a macaque to carry out neural modulation. The electrode array relied on cytochrome oxidase (CO) staining to complete the topological functional zoning of the visual cortex. PEDOT:PSS composite coating modified platinum electrodes were prepared on the surface of the electrodes to achieve long-term electrical stability and implantable biocompatibility encapsulation.

[0095] Phase 1: Image Input Preprocessing and DKL Color Space Parameter Calculation

[0096] After the camera captures the original apple image, it sequentially performs three standardized preprocessing steps: white balance, demosaicing, and gamma correction, outputting a standard sRGB format image with a baseline value of R=255, G=0, and B=0 for the sample red pixels. To match the natural visual neural response mechanism of the primate retina-lateral geniculate body, the sRGB pixel values ​​need to be converted to the Derrington-Krauskopf-Lennie (DKL) antagonistic color space.

[0097] The DKL color space is constructed based on the color antagonistic neural pathways of the two major visual centers: retinal ganglion cells and the lateral geniculate nucleus (LGN). The spatial coordinate axis can directly correspond to two core antagonistic conduction pathways: the LM red-green antagonistic pathway and the S-(L+M) blue-yellow antagonistic pathway, which fits the coding logic of living visual perception.

[0098] The conversion from RGB color to DKL color uses the LMS view frustum response space as an intermediate transition space. The standardized conversion process and axis definition are as follows:

[0099] Step 1: RGB to LMS frustum quantum capture rate: Based on the frustum relative sensitivity conversion matrix after factory calibration of the CMOS / image sensor, the mapping and conversion of pixel grayscale values ​​to L (long-wavelength frustum), M (medium-wavelength frustum), and S (short-wavelength frustum) frustum quantum capture rates are completed.

[0100] Step 2: Convert LMS to DKL three-dimensional feature coordinates (τ, δ, ε): The three coordinate axes correspond to independent visual encoding dimensions: ① τ Brightness axis: Integrates L+M cone response signals, corresponding to visual brightness information. The signal is mainly transmitted through the lateral geniculate macrocell pathway; ② δ Red-green antagonism axis: Characterizes red-green antagonistic response. Positive values ​​within the axis induce subjective red perception, and negative values ​​induce subjective green perception; ③ ε Blue-yellow antagonism axis: Characterizes blue-yellow antagonistic response. Positive values ​​within the axis induce subjective blue perception, and negative values ​​induce subjective yellow perception.

[0101] In this experiment, the visual stimulus of the red target, after DKL space calculation, showed clear characteristics: the δ axis showed a high positive value component (sample δ≈+0.8), the ε axis component was infinitely close to 0, and there was no blue-yellow antagonistic color response.

[0102] Based on existing publicly available colorimetric benchmark datasets: the standard red stimulus has coordinates [0.416, −0.909, 0.000] in the cone luminance antagonism space, with the three-dimensional components corresponding to the L+M luminance channel, the LM red-green antagonism channel, and the S-(L+M) blue-yellow antagonism channel, respectively. It should be noted that −0.909 in the LM channel is a custom sign rule for the coordinate system; the sign only represents the antagonistic polarity, and the absolute value characterizes the high-intensity modulation effect of red on the red-green antagonistic neural pathway, consistent with the conclusion of high positive values ​​in the δ-axis perception of the DKL space in this experiment.

[0103] Analysis of the DKL isoluminance chromatogram rings shows that standard red falls within the 0°~180° range of the LM principal axis, corresponding to a polar angle θ≈0° (equivalent to 360°); color saturation is quantified and characterized by the polar radius r. The final standardized output is the {LUM, r, θ} triplet DKL feature parameters, which are then sent to the backend electrical stimulation algorithm engine. The parameter definitions are as follows:

[0104] - LUM: Z-axis brightness coordinate in DKL space, independently representing the brightness and non-chromatic visual information of an image;

[0105] - r: Color polar radius, representing the antagonistic modulation amplitude of pixel color relative to the energy-adaptive white point (space origin) such as DKL, which is equivalent to the color saturation perceived by the human eye;

[0106] -θ: Polar angle of the isoluminance plane, representing the angle by which the chromaticity signal deviates from the pure luminance axis. θ=0° represents pure isoluminance chromaticity stimulation.

[0107] Phase Two: The algorithm engine constructs the spatiotemporal stimulus matrix and electrical control parameters.

[0108] The waveform scheduler receives the DKL triplet feature parameters {LUM, r, θ} sent by the front end, as well as the LM channel-specific weight coefficient w that matches the CO functional block of the V1 region. It then solves the multidimensional temporal electrical stimulation matrix I(t) of the cortical site using a spatiotemporal tensor fusion algorithm.

[0109] For targeted electrical stimulation of the red-specific CO functional block in region V1, the temporal stimulation matrix I(t) must satisfy three constraints: charge quantity, waveform temporal domain, and neural perception.

[0110] 1. Charge Constraint: The charge amount of a single-cycle stimulation is higher than the activation threshold of the visual cortex nerves, while being strictly lower than the upper limit of the safe charge amount for in vivo implantation. For details of the safe threshold standard, please refer to the experimental specifications in Phase 5.

[0111] 2. Temporal waveform constraints: The square wave pulse stimulation method is abandoned and a gradual envelope waveform is adopted. The parameters are set as a linear rising edge of 5~10μs and a steady plateau region of 100~300μs to weaken the high-frequency harmonics of the pulse and reduce the risk of ectopic activation of non-target nerves in the cortex.

[0112] 3. Visual perception constraints: Referring to behavioral calibration data within 5° of the central visual field of primate V1 area: the diameter range of the photic hallucination induced by single-point microelectrode electrical stimulation is 9~26 arcminutes, and the effective perceptual contrast range is 2.6%~10%. This set of quantitative data directly binds to the bidirectional mapping relationship between the amplitude of the stimulation matrix and the spatial resolution of the cortical stimulation point.

[0113] For the pure red antagonistic stimulus at θ=0° in this experiment, the algorithm retrieves the global electrode weight matrix calibrated in the pre-experiment and extracts a dedicated one-dimensional electrode weight vector adapted to the 0° color antagonism. The experiment configures a dimensionless electrode allocation weight matrix W, satisfying the normalized summation of the active stimulation electrode weight coefficients (the sum of the coefficients is 1). Simultaneously, a shielded loop electrode is configured to ensure that the total output current of the shielded electrode is equal in absolute value and opposite in polarity to the active electrode current, achieving zero net charge at the initial cortical implantation site and avoiding electrode corrosion and neuroinflammatory damage caused by charge accumulation.

[0114] Phase 3: ASIC Hardware Configuration and Electrode Topology Activation

[0115] The core hardware of this visual reconstruction architecture is a high-voltage compliant dedicated neural stimulation ASIC chip fabricated using standard CMOS technology. The chip relies on a three-phase clamping waveform driving model to target and regulate the V1 region CO Blob functional area, precisely regulate the red-green antagonistic neural circuit to complete color vision activation, suppress ectopic neural responses to the greatest extent, and reduce optical illusions and color interference.

[0116] 3.1 CMOS Programmable Switch Matrix

[0117] This system uses an integrated CMOS programmable switch matrix to replace the traditional discrete mechanical switch array, enabling dynamic routing coupling between the constant current drive path and any microelectrode in the array, and switching the electrode operating mode as needed. In this red light sensing experiment, the CMOS switch matrix embeds three independent, isolated logic drive links, each performing its specific function to control the electric field and isolate stray emissions.

[0118] Link A (Active Red Stimulation Pathway): Binds to the dedicated active electrode E1 within the CO Blob functional region, serving as the core site for injecting red antagonistic signal current.

[0119] Link B (electric field shielding return path): It is bound to the shielding electrodes E2 and E3 arranged adjacent to E1. It is used to absorb the lateral diffusion current of the cortex, compress the electric field diffusion range, and create a targeted stimulation electric field with steep boundaries and strong locality.

[0120] Link C (Interblob Isolation Pathway): The entire area of ​​the contour electrode E4 in the interblob region outside the Blob is set to a high-resistance state, completely disconnecting the bus drive link, so that this type of electrode does not participate in the pulse stimulation, avoiding interference with the dedicated neural pathway of the shape contour of the interblob region of the cortex, and eliminating the illusion of contour perception coupled with stray light.

[0121] The aforementioned zoned electrode routing strategy is supported by mature neurobiological theories: in the primate V1 cortex, the Blob region, which is highly enriched by CO staining, is dedicated to processing visual surface attribute information such as color and brightness; the Interblob region is dedicated to processing shape and structural information such as object edges and spatial contours. In vivo in vivo light imaging studies have confirmed that the spatial distribution of color-selective neuron clusters in the V1 cortex highly overlaps with CO Blob patches, and zoned stimulation can achieve specific activation of the color vision pathway.

[0122] Before the formal initiation of DKL color electrical stimulation, the system must complete a round of induced closed-loop functional mapping. This process does not rely on in vitro histological staining or in vitro optical imaging equipment; it only relies on the integrated stimulation-recording duplex capability of the implanted electrode array to in-situ analyze the V1 cortical CO Blob functional partitions, automatically classify the functions of the entire electrode domain, and generate a hardware-specific routing table, providing hardware topology basis for subsequent precise DKL color current injection. The entire mapping is completed collaboratively by on-chip weak electrical signal transceiver links and software-layer unsupervised clustering algorithms. The following section uses a 320-channel flexible rectangular cortical electrode array with a center-to-center spacing of 100 μm as a case study to refine the closed-loop mapping process, parameter standards, and iterative optimization logic.

[0123] To prevent damage from DC charge accumulation at the electrode-tissue interface (ETI), a symmetrical biphase constant current pulse pattern is uniformly used for the full-domain mapping and detection pulses. An on-chip residual voltage monitor (RVM) is also incorporated; after each pulse cycle, a closed-loop zeroing of the interface potential is performed immediately to maintain electrical steady-state at the interface. The standardized detection parameters for closed-loop mapping are shown in the table below:

[0124] Table 1: Standardized Detection Parameters for Closed-Loop Mapping

[0125] Parameters Setting value in accordance with Detecting current amplitude 0.5–1.0 μA This is far below the V1 minimum optic phantom threshold current reported in the literature (1.9–5 μA, Schmidt et al., 1996). Monophasic pulse width 50 μs Minimal injected charge (50–100 pC / phase), interface voltage perturbation < ±20 mV Interphase delay 10 μs Avoid cross-coupling Pulse repetition frequency 10 Hz Each channel can continuously record for 1 second to obtain the average of 10 evoked potentials. Two-phase charge balance RVM Active Clamping After each pulse, Phase 3 constant voltage clamping is performed to 0 mV to ensure zero net charge.

[0126] The full-domain channel uses a time-division multiplexing (TDM) row and column scanning mode for point-by-point detection by default. To reduce the mapping time, the on-chip multi-channel parallel acquisition resources of the ASIC can be called to simultaneously detect non-adjacent electrode points in space. The total time for full-domain closed-loop mapping can be reduced to less than 5 minutes.

[0127] 3.2 Multidimensional Electrophysiological Feature Extraction and Blob Probability Mask Mapping

[0128] During the full-domain pulse scan, the on-chip acquisition link synchronously acquires the local field potential (LFP) at a sampling rate of ≥2kHz. Three independent electrophysiological features are extracted for each probe electrode to construct a standardized electrode feature vector: f_i =(Z, A, La).

[0129] 3.2.1 Interface impedance mode characteristics (|Z| 1 kHz)

[0130] By selecting the midpoint of the probe pulse anode phase, a high-precision ADC synchronously acquires the real-time voltage at the electrode interface. Combined with the calibrated injection current value, the impedance modulus at 1kHz of the single-channel electrode-brain tissue interface is calculated. Based on the cortical regional conductivity model, the CO Blob region has higher vascular density and larger extracellular space volume, resulting in a significantly lower equivalent interface impedance than the surrounding Interblob region. Combined with the 100 μm electrode spacing of this array, the impedance spatial gradient can accurately determine the physical boundary of the Blob functional area.

[0131] 3.2.2 Peak-to-peak amplitude of induced local field potential

[0132] LFP signals evoked by cross-site coupling in each channel were extracted, and the physiological latency time window was defined as 2–10 ms. The peak-to-peak potential amplitudes of the P1-N1 signals were calculated. The neural circuit characteristics showed that the recurrent excitatory synaptic connections within the Blob were more dense, and it also received high-density projection input from the lateral geniculate body of the thalamus. The evoked potential amplitudes were significantly higher than those in the Interblob circuit. This pattern has been verified in in vivo ICMS-LFP spatial propagation studies (Butovas & Schwarz, 2003).

[0133] 3.2.3 Logarithmic slope of δ-α low-frequency power spectral density

[0134] The native LFP signal of each electrode in the 1s pulse-free resting state was collected, and the logarithmic slope of the power spectral density in the 1–12 Hz δ-α band was calculated as a feature for differentiating resting electrical activity in cortical regions.

[0135] After acquiring the full-domain 3D feature data, Gaussian kernel spatial smoothing noise reduction is applied with smoothing kernel parameter σ = 50 μm. The half-electrode spacing is matched to remove single-channel random measurement noise and restore the continuous contour of the cortical functional areas. Then, the Otsu adaptive threshold algorithm and neighborhood pixel continuity constraint are fused to generate the CO Blob binary functional mask M(x,y). The mask value rules are defined as follows: M(x,y)=1 represents that the electrode falls into the Blob color vision functional area, and M(x,y)=0 represents that the electrode falls into the Interblob contour functional area.

[0136] Closed-loop iterative optimization mechanism: After the first round of mask mapping, based on the existing geometric partition prior information, boundary iterative confirmation detection is carried out: only for blurred points on the edge of the blob and adjacent interblob electrodes, the detection pulse frequency is increased to 20 Hz, the feature vector is updated by secondary sampling, and the mask is iteratively updated until the boundary information entropy converges and stabilizes. After iterative optimization, the accuracy of blob boundary determination is improved from the initial ±30 μm to ±10 μm, which significantly improves the accuracy of electrode classification.

[0137] 3.3 Electrode Function Classification and Optimal Active Electrode Selection Algorithm

[0138] Based on the converged binary functional mask, the functional logic labeling of 320 full-channel electrodes is completed. The core optimization logic is to combine 8-neighbor connected domain analysis, take maximizing the number of available shielded electrodes as the selection criterion, and select the optimal active stimulation electrode within a single Blob region.

[0139] 3.3.1 Initial Logic Marking of Electrodes

[0140] If the mask coordinates M(i,j) = 1: the electrode (i,j) is included in the electrode pool within the Blob region, and can be selected as an active electrode or a shielding electrode;

[0141] If the mask coordinates M(i,j) = 0: the electrode (i,j) is directly marked as an isolation electrode, and the CMOS switch matrix directly disconnects the bus link.

[0142] 3.3.2 Active Electrode Screening and Intrinsic Shielding Ring Construction

[0143] Blob connectivity segmentation: Using the 8-neighborhood connectivity determination rule, spatially connected Blob electrode clusters are clustered to obtain a set of independent functional blocks {C1, C2, …, C_K}, with each set corresponding to an independent CO Blob color vision functional patch.

[0144] Calculation of shielding adaptation coefficient S(e): For any electrode e in the connected domain C_k, define a set of neighboring electrodes N(e) with a radius ≤200 μm; the number of Blob electrodes in the same block within the set (excluding electrode e itself) is defined as the shielding adaptation coefficient S(e) of that electrode. The larger the coefficient, the more internal shielding electrodes can be configured, and the better the electric field confinement effect.

[0145] Optimal active electrode selection rule: Select the electrode with the maximum value of S(e) in a single connected domain as the core active electrode; if the S(e) values ​​of multiple electrodes are equal, select the electrode closest to the geometric centroid of the connected domain to ensure that the current injection site is located in the core of the functional area and improve the uniformity of stimulation across the entire domain.

[0146] Electrode secondary marking and division: After determining the optimal active electrode e*, take the intersection electrode of the neighborhood set N(e*) and the current connected domain and designate it as a dedicated shielding electrode. All shielding electrodes are limited to the Blob region to prevent current leakage from interfering with the Interblob contour path. The remaining unselected electrodes in the Blob region are uniformly marked as backup active electrodes, sorted in descending order according to the S(e) value, and stored in the on-chip electrode priority scheduling table.

[0147] Extreme Block Adaptive Control Strategy: If the maximum S(e) of a single Blob connected region is less than 4, the intrinsic electrode within the region cannot construct a complete shielding ring, and the system adopts hierarchical adaptive control:

[0148] Strategy A: Expand the internal shielding neighborhood, extending the shielding radius to 300 μm, still limited to electrode network shielding within the Blob area; if network formation fails, execute Strategy B.

[0149] Strategy B: Enable temporary inter-spot auxiliary shielding. Select 2-4 Interblob electrodes within the vicinity of active electrode 8 to temporarily form a network shield. These electrodes normally maintain an isolated high-resistivity state. Only the stimulation phase temporarily conducts backflow. It recovers immediately after stimulation ends. This condition is written into the abnormal flag bit of the routing table for subsequent fine-tuning of stimulation parameters.

[0150] Strategy C: Minimal patch rejection strategy. If the maximum S(e) of a patch is less than 2, the patch is determined to be an imaging artifact or a fragmented invalid blob. The entire patch is marked as an isolation electrode and does not participate in color vision stimulation.

[0151] 3.3.3 Automatic generation of global hardware routing table

[0152] After the closed-loop mapping iteration converges, the off-chip FPGA main control unit coordinates the on-chip storage unit to generate a 320-row × 8-field underlying hardware-specific routing table. Each channel electrode corresponds to one routing configuration entry, directly defining the CMOS switch operating state. The meanings of the routing table fields are as follows:

[0153] Table 2: Routing Table Fields

[0154] Fields illustrate Electrode ID Unique hardware address (0–319) Physical (row,col) physical coordinates Mask Label 0 = Isolation, 1 = Active, 2 = Shielding, 3 = Backup Active Connected Active If it is a shielded electrode, specify the ID of the active electrode it belongs to (otherwise, null). Priority If it is a backup active, record the activation priority within this Blob (1 is the highest). Shielding Count The actual number of shielding ring electrodes corresponding to this electrode (if active). Switch State Initial CMOS switch encoding (corresponding to high impedance / cathode drive / anode drive) Anomaly Flag Set to 1 if Interblob auxiliary masking is enabled, otherwise set to 0.

[0155] For clarity, assume that the center of a certain coverage area of ​​this 320-electrode array contains a relatively complete COBlob connected region {C1}, consisting of 9 electrodes, arranged as follows: Figure 4 .

[0156] 1. Step 1 Scan Mask Determination: After global pulse detection, the algorithm generates a convergent binary mask and determines that the nine-grid global electrodes belong to the connected region {C1}, which is an electrode within the Blob functional area.

[0157] 2. Step 2: Shielding Adaptability Coefficient Calculation and Optimization: Based on Chebyshev neighborhood distance calculation, the S-value is calculated as follows: Center electrode E5 has 4 effective neighboring Blob electrodes, S=4; Edge electrode E4 has 3 effective neighboring electrodes, S=3; Corner electrode E1 has 2 effective neighboring electrodes, S=2. Finally, the centroid electrode E5 is selected as the dedicated active electrode for red light; Corner electrodes E1, E3, E7, and E9 are adjacent to the Interblob area, and are directly marked as isolation electrodes and their drive links are disconnected.

[0158] 3. Step 3 Internal Source Shielding Ring Network: Select four internal source electrodes E2, E4, E6, and E8 to form a closed-loop surrounding shielded return ring. All shielding electrodes are located within the Blob area, with no cross-area leakage.

[0159] 4. Step 4 Hardware Routing Parameter Configuration: The routing table is configured with E5 in cathode current drive mode, and all four shielded electrodes are uniformly configured with anode return current drive and bound to their respective E5; the isolation electrodes are set to a high-resistance state throughout. This topology ensures that the first phase current of red light stimulation is injected from a single point on E5, with an equal amount of return current on the surrounding shielded electrodes, and the electric field potential well is completely confined within the Blob boundary, achieving atomic-level precise activation of the color vision neural pathway.

[0160] After the induced closed-loop mapping is completed, the solidified hardware routing table is written into the on-chip state machine lookup table of the CMOS switch matrix as a long-term fixed hardware configuration, eliminating the need for repeated mapping and calibration for regular visual frame stimuli. Subsequently, after each frame image from the camera undergoes stage one DKL calculation and stage two algorithm weighting, the waveform scheduler directly retrieves the routing table topology information and uses directional drive to bind the active and shielded electrodes to complete color vision electrical stimulation.

[0161] Meanwhile, the on-chip RVM module continuously monitors the interface electrical parameters, acquiring real-time residual polarization voltage and interface impedance values ​​of the active electrode ETI. When the impedance drift of the working electrode E5 exceeds 30% and the interface safety charge window deteriorates, the on-chip state machine automatically triggers a seamless hot-switching: it retrieves the priority sequence of the backup electrode from the routing table, instantly switches the cathode drive to the highest priority backup electrode, and simultaneously reconstructs the surrounding shielding loop. The entire switching process takes less than 100 μs, ensuring uninterrupted visual stimulation throughout the entire process and providing long-term stability for red light target stimulation.

[0162] The RVM hardware port is directly connected to the CMOS switch matrix, which can time-division multiplex each microelectrode in the array. After the second phase of each stimulation is completed, a microsecond-level high-precision sampling window is opened. Relying on the on-chip high-resolution ADC, the polarization analog potential of the electrode interface is converted into a digital quantity, providing reference electrical parameters for cross-frame charge closed-loop correction.

[0163] Even when a theoretically symmetrical biphase balanced pulse is delivered, residual DC bias will still remain at the electrode interface due to the Faraday impedance charge leakage and the polarization characteristics of the interface material. Long-term accumulation of this bias will induce electrode corrosion and colloid inflammatory hyperplasia. By relying on real-time feedback of the residual potential value via RVM, and performing closed-loop control with phase pre-compensation for the next pulse, the DC bias at the interface can be completely eliminated, significantly improving the safety of long-term implantation stimulation and device lifespan.

[0164] Phase Four: Three-Phase Waveform Modulation and "No-Stirring" Color Reconstruction Closed Loop

[0165] The system employs an arbitrary waveform generator (AWG) to decompose the timing stimulus signal (I(t)) into microsecond-level fine time steps, generating gently rising / stepped waveforms to avoid electrical interference from steep-edge pulses. The entire three-phase stimulus waveform is executed strictly according to a closed-loop pipeline timing sequence. The operating mechanism of each phase is as follows: Figure 5 .

[0166] Phase 1: Active stimulation and spatial electric field clamping (cathode injection + spatial constraint)

[0167] Time window: Total duration 190 μs, including 10 μs cathode ramp-up section + 180 μs cathode steady-state constant current platform section.

[0168] Electrical operating status: Active electrode E1 is connected to the controlled cathode constant current source according to preset weights, while shielded electrodes E2 and E3 are connected to the anode constant current circuit. The stimulation current gradually rises through a 10 μs ramp waveform, abandoning the steep square wave shape, effectively avoiding interference from high-frequency components introduced by the square wave on the large cell pathway, while reducing the risk of damage to nerve tissue by high-order harmonics.

[0169] Physical effects: A negative potential well with a very steep gradient is formed within the CO Blob functional region; the shielding ring electrode synchronously absorbs an equal amount of reverse current, achieving strong spatial confinement of the stimulation electric field.

[0170] Phase 2: Active charge recovery and cross-frame adaptive modulation (anodic recharging + inter-frame dynamic compensation)

[0171] Time window: 190 μs~230 μs interval, pulse polarity is reversed overall, asymmetric adaptive waveform is used, phase duration and anode output current amplitude are independent of the phase 1 cathode stimulation parameters, and there is no fixed matching ratio.

[0172] Closed-loop control logic: If the residual voltage monitor (RVM) detects an accumulated negative polarity potential of −35μV at the interface after the stimulation of the previous frame, the algorithm engine will perform dynamic compensation: increase the total duration of Phase 2 of this frame by 0.5% (about 1 μs), or fine-tune the amplitude of the anode pulse output of the shielded electrode, and achieve cross-frame charge dynamic balance by injecting positive charge in advance.

[0173] Phase 3: Constant potential clamping and microsecond-level potential clearing

[0174] Time window: Start immediately within 100 ns after the end of Phase 2, with a total duration of ≤10 μs.

[0175] Operating Logic: The neurostimulation chip utilizes a dedicated low-resistance discharge circuit. The system combines real-time potential drift data acquired by the RVM to calculate the interface reference constant potential, actively neutralizing the residual nonlinear charge in the double-layer capacitance (Cdl) at the electrode interface. Unlike passive methods that rely on natural discharge via RC circuits, this active discharge mechanism completely eliminates stimulation artifacts such as pulse polarization tails, restoring the electrode-tissue interface to an electrochemically neutral state and establishing a stable tissue impedance baseline for subsequent continuous pulse sequences.

[0176] After the three-phase waveform timing sequence is completed, the excited color-sensing neurons in the CO Blob region undergo depolarization and generate an action potential sequence under the action of the progressive cathode current in Phase 1. This electrical activity can drive the synchronous firing of color-selective neuronal clusters in the Blob region of layers 2 / 3 of the V1 cortex, thereby inducing two types of characteristic responses: one is the directly measurable local field potential of neurophysiology, and the other is the color-dependent BOLD-fMRI metabolic signal. The former characterizes the instantaneous electrical activity of neurons, while the latter reflects the metabolic changes in brain tissue coupled with neural activity.

[0177] Neurons in the CO Blob region highly express cytochrome oxidase (COX), exhibiting significantly higher oxidative metabolic capacity than those in the surrounding Interblob region. Furthermore, Blob neurons naturally possess a preferential response to color stimuli. The rich enzyme system and dense capillary network in this region provide ample energy and blood flow for high-frequency color-related neural activity, forming a good biological fit with the system's mechanism of targeted injection of focused current into the CO Blob to induce high-intensity neuronal firing.

[0178] Activated Blob neurons project across cortical layers, transmitting signals to the V2 cortical striate area. These signals then travel through V4, the core brain region responsible for color constancy calculations, and finally to higher visual centers such as the infratemporal cortex (IT), integrating to form a subjective visual hallucination with a clear red color perception attribute. The spatial location of this hallucination is determined by the topological mapping of the visual cortex, and the red color perception originates from the specific activation of the LM red-green antagonistic pathway, rather than a generalized response of the brightness pathway.

[0179] Based on the findings of Schmidt et al.'s experiments on intracortical microcurrent stimulation (ICMS) in blind individuals, optic hallucinations induced by multi-electrode synchronous stimulation exhibit coplanarity and relative positional constancy, while single-electrode stimulation near the threshold can induce color optic hallucinations. This system utilizes CO Blob targeted stimulation to precisely activate the color vision module, Interblob electrical isolation to cut off shape and motion information processing pathways, and RVM and constant potential clamping mechanisms to eliminate polarization interference, achieving pure chromaticity signal transmission from a physiological perspective and completely avoiding stray color noise.

[0180] Based on the above architecture, users can obtain a stable and pure visual perception effect:

[0181] Spatial characteristics: Optical illusions appear at the corresponding positions in the field of view. The diameter of the light spot within the central field of view is 9 to 26 arcminutes, and the specific size is determined by the field of view magnification factor corresponding to the electrode position.

[0182] Color perception: It produces a highly saturated, pure red light perception without black and white spots or other color interference. The core reason is that the stimulation target is precisely located on the color vision functional column (CO Blob), specifically activating the LM red-green antagonistic neural pathway.

[0183] Temporal performance: The AWG ramp waveform shapes the pulse edge, making the onset and fading of photopsia smooth and natural, without obvious flicker; the stimulation duration is about 200 ms, which, after superimposed with the inherent 47~55 ms signal delay of the cortex, forms a continuous and smooth red visual experience.

[0184] By using a high-density electrode array to perform spatiotemporal combined electrical stimulation, multiple independent "red visual pixels" can be spliced ​​and fused together to ultimately reconstruct a complete color image, providing technical support for visually impaired people to rebuild color vision.

[0185] Similar to the above embodiments, the present invention provides a method for reconstructing color vision function of the visual cortex based on three-phase clamping waveforms.

[0186] The following specific embodiments are provided in conjunction with the accompanying drawings:

[0187] like Figure 6This document illustrates a flowchart of a method for reconstructing visual cortex color vision function based on a three-phase clamping waveform, as described in an embodiment of the present invention. Applied to the three-phase clamping waveform-based visual cortex color vision function reconstruction system of the above embodiment, the method includes:

[0188] Step S1: Acquire the collected visual image and the residual polarization voltage detected by the residual voltage monitor;

[0189] Step S2: Based on the acquired visual image and the residual polarization voltage detected by the residual voltage monitor, a color vision control command is generated to control the CMOS switch matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load, thereby realizing the reconstruction of color vision function in the visual cortex; wherein, the three-phase closed-loop stimulation waveform is an asymmetric waveform envelope.

[0190] Since the implementation principle of the visual cortex color vision function reconstruction method based on three-phase clamping waveform has been described in the foregoing embodiments, it will not be repeated here.

[0191] The visual cortex color vision function reconstruction system based on three-phase clamping waveforms provided in this invention can be implemented on the terminal side or the server side. For the hardware structure of the electronic terminal, please refer to [link to relevant documentation]. Figure 7 This is a schematic diagram of an optional hardware structure of an electronic terminal 1000 provided in an embodiment of the present invention. The terminal 1000 can be a mobile phone, computer device, tablet device, personal digital processing device, factory back-end processing device, etc. The terminal 1000 includes: at least one processor 1001, a memory 1002, at least one network interface 10010, and a user interface 1009. The various components in the device are coupled together through a bus system 1005. It is understood that the bus system 1005 is used to realize the connection and communication between these components. In addition to a data bus, the bus system 1005 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 7 All buses are labeled as bus systems.

[0192] The user interface 1009 may include a monitor, keyboard, mouse, trackball, clicker, button, touchpad, or touch screen.

[0193] It is understood that memory 1002 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM) or programmable read-only memory (PROM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM) and synchronous static random access memory (SSRAM). The memories described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable categories of memory.

[0194] In this embodiment of the invention, the memory 1002 is used to store various types of data to support the operation of the terminal 1000. Examples of this data include: any executable program for operation on the terminal 1000, such as the operating system 10021 and application program 10022; the operating system 10021 contains various system programs, such as the framework layer, core library layer, driver layer, etc., for implementing various basic services and handling hardware-based tasks. The application program 10022 may contain various applications, such as a media player, browser, etc., for implementing various application services. The visual cortex color vision function reconstruction system based on three-phase clamping waveform provided in this embodiment of the invention can be included in the application program 10022.

[0195] The methods disclosed in the above embodiments of the present invention can be applied to or implemented by the processor 1001. The processor 1001 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 1001 or by instructions in the form of software. The processor 1001 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 1001 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor 1001 may be a microprocessor or any conventional processor, etc. The steps of the accessory optimization method provided in the embodiments of the present invention can be directly reflected as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium, which is located in a memory. The processor reads the information in the memory and combines it with its hardware to complete the steps of the aforementioned method.

[0196] In an exemplary embodiment, the terminal 1000 may be used to execute the aforementioned method by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), or complex programmable logic devices (CPLDs).

[0197] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented using computer program-related hardware. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0198] In the embodiments provided in this application, the computer-readable and writable storage medium may include read-only memory, random access memory, EEPROM, CD-ROM or other optical disc storage devices, disk storage devices or other magnetic storage devices, flash memory, USB flash drive, portable hard drive, or any other medium capable of storing desired program code in the form of instructions or data structures and accessible by a computer. Additionally, any connection may be appropriately referred to as a computer-readable medium. For example, if instructions are transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. However, it should be understood that computer-readable and writable storage media and data storage media do not include connections, carrier waves, signals, or other transient media, but are intended for non-transient, tangible storage media. The disks and optical discs used in the application include compact optical discs (CDs), laser optical discs, optical discs, digital multifunction optical discs (DVDs), floppy disks, and Blu-ray discs, where disks typically copy data magnetically, while optical discs use lasers to copy data optically.

[0199] Compared with the prior art, the present invention has the following advantages:

[0200] I. Achieving Iterative Upgrades of Stimulation Waveforms to Achieve Physical-Level Zero Charge at the Interface. Traditional implanted stimulators only use cathode injection + anode recovery to fix a biphasic square wave, which cannot eliminate residual polarization potential at the interface, and the steep-edge square wave is prone to inducing high-frequency harmonic damage in brain tissue. This invention introduces a multi-timestep analytical AWG arbitrary waveform architecture, decoupling the three-phase timing of injection, charge recovery, and active potential clamping, which can generate ramp and step composite asymmetric waveform envelopes; combined with a dedicated clamping design for the third phase, the residual potential at the electrode interface is controlled within the academically recognized ±50mV non-destructive safety window, achieving complete physical-level zeroing of nonlinear interface charge, which is different from the traditional waveform approximate charge balance mode.

[0201] II. Achieving Blob-Targeted Isolation Spatial Clamping Based on CMOS Switch Matrix Topology Networking. This invention abandons the divergent unipolar stimulation and fixed hard-wired switching mode, and constructs a software-defined dynamic multipolar stimulation topology: through dynamic coupling of constant current driver and electrode cluster by CMOS switch matrix, the active core electrode at the center of the Blob executes the DKL color pathway stimulation command, and the shielding ring electrode within the plaque synchronously and equally flows back current according to Kirchhoff's current law, so that the electric field decays exponentially across the shielding ring, completely confining the activated tissue volume VTA within the CO Blob color vision area, and completely isolating the surrounding Interblob contour perception area, thus eliminating cross-regional neural crosstalk at the spatial level.

[0202] Third, the invention employs an RVM closed-loop clamping architecture to address the pain point of interface nonlinear discharge hysteresis. This invention sets up a hardware-level ultra-fast closed-loop control link. After the active charge recovery and cross-frame adaptive modulation stages, the RVM instantaneously acquires the interface polarization residual potential with microvolt precision. It abandons the traditional passive grounding discharge mode, using a digital state machine combined with the sampled data to close a low-resistance discharge loop, driving the loop to a reference constant potential. This can accelerate the neutralization of nonlinear residual charge in the double-layer capacitance within 10μs, precisely clamping the interface potential to the zero potential range, thus solving the millisecond-level hysteresis and polarization tailing problems of traditional discharge.

[0203] IV. Equipped with a cross-frame adaptive modulation mechanism to eliminate color distortion and ensure color fidelity. Relying on the full-time monitoring capability of RVM, the system can identify inter-frame polarization drift in real time and adaptively adjust the amplitude and pulse width of the anode pulse in the next frame to prevent long-term charge accumulation. At the same time, the polarization tail of the preceding pulse is completely eliminated, providing an electrically neutral impedance baseline for subsequent color stimuli. This ensures that the DKL antagonistic color neural code is accurately converted into neuronal membrane potential signals, avoiding hue mixing and color distortion problems, and achieving high-fidelity pure color light illusion reconstruction.

[0204] V. ASIC Hardware-Level KCL Current Closed-Loop Verification to Lock in the Electrochemical Safety Boundaries of Brain Tissue. The chip of this invention employs a calibrated dual-current driver paired coupling architecture at its core. Relying on flexible routing via a CMOS matrix, the shielding loop return current can be mirrored and synchronized with the active core injection current at the nanosecond level, forming a local closed current loop and eliminating stray currents across the entire domain caused by leakage from the distal reference electrode. Combined with a three-phase clamping potential zeroing mechanism, the interface polarization voltage is strictly controlled within the electrochemical safety window, avoiding tissue pH drift, electrode corrosion, and excitotoxicity, thus meeting the safety requirements for long-term implantation.

[0205] In summary, the visual cortex color vision reconstruction system, method, and terminal based on three-phase clamped waveform of the present invention are equipped with an electrode array, which includes an active core electrode, a shielding electrode, and an isolation electrode. The active core electrode is placed in the central region of the target area of ​​visual cortex color vision function, the shielding electrode is placed adjacent to the active core electrode inside the target area of ​​visual cortex color vision function, forming an electric field shielding ring, and the isolation electrode is placed in the interval area of ​​visual cortex color vision function area; the physiological tissue between the active core electrode and the shielding electrode constitutes the stimulation load. The system integrates a CMOS switch matrix, a residual voltage monitor, and a central processing control module: the switch matrix drives the active core electrode and the shielding electrode respectively through independent constant current drivers; the residual voltage monitor collects the residual polarization voltage of the stimulation load in real time; the central processing control module combines visual image information and polarization feedback voltage to issue commands to control the switch matrix to output an asymmetric envelope three-phase closed-loop stimulation waveform, thereby completing accurate color vision reconstruction. This invention integrates AWG arbitrary waveform generation, CMOS programmable gating, and RVM residual voltage monitoring hardware to build a four-dimensional dynamic charge balance system. Relying on spatial electric field potential well constraint and microsecond-level active clamping cross-frame modulation, it constructs an electrochemically neutral tissue impedance baseline, overcoming the technical shortcomings of traditional electrostimulation's spatiotemporal separation. Therefore, this invention effectively overcomes the various shortcomings of existing technologies and has high industrial application value.

[0206] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A visual cortex color vision function reconstruction system based on three-phase clamped waveforms, characterized in that, The system includes: An electrode array includes an active core electrode, multiple shielding electrodes, and multiple isolation electrodes. The active core electrode is located in the central region of the color vision target area of ​​the visual cortex. The multiple shielding electrodes are adjacent to the active core electrode and located in the color vision target area of ​​the visual cortex, forming a shielding ring. The multiple isolation electrodes are located in the interval regions of the color vision target area of ​​the visual cortex. The physiological tissue between the active core electrode and each shielding electrode constitutes the stimulation load. The CMOS switch matrix includes: a first constant current driver connected to the active core electrode and a second constant current driver each connected to a shield electrode; A residual voltage monitor, connected to the active nuclear electrode and the shielding electrode, is used to sample the residual polarization voltage of each stimulus load in real time. The central processing and control module, connected to the residual voltage monitor, is used to generate color vision control commands based on the acquired visual images and residual polarization voltage, and control the CMOS switch matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load to realize the reconstruction of color vision function in the visual cortex; wherein, the three-phase closed-loop stimulation waveform is an asymmetric waveform envelope.

2. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 1, characterized in that, The three-phase closed-loop stimulation waveform corresponds to three independent stages executed sequentially, including: active stimulation and spatial clamping stage, active charge recovery and cross-frame adaptive modulation stage, and constant potential clamping and zeroing stage; wherein, the waveforms of the active stimulation and spatial clamping stage and the active charge recovery and cross-frame adaptive modulation stage contain asymmetric waveform envelopes with ramp and / or step features.

3. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 2, characterized in that, The method of generating color vision control commands based on acquired visual images and residual polarization voltage, and controlling the CMOS switching matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load, includes: The acquired visual image frames are mapped to the DKL color space, and the DKL logical parameters are extracted. A multidimensional stimulation matrix is ​​generated by the waveform scheduler based on the DKL logic parameters and the residual polarization voltage measured after the end of the previous frame stimulation. The stimulation matrix is ​​decomposed into multiple time steps by arbitrary waveform generation logic to generate color vision control instructions, and the CMOS switching matrix is ​​controlled by a hardware state machine to execute a three-phase closed-loop stimulation waveform on the stimulation load.

4. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 3, characterized in that, When generating the multidimensional stimulation matrix, the central processing and control module combines the residual polarization voltage measured after the end of the previous frame stimulation to locally correct the parameters of the active charge recovery and cross-frame adaptive modulation stages in the multidimensional stimulation matrix; wherein, the parameters include: current amplitude or duration.

5. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 4, characterized in that, The process of controlling the CMOS switching matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load via a hardware state machine includes: During the active stimulation and spatial clamping phase, the hardware state machine controls the CMOS switch matrix to connect the active nuclear electrode to the constant current source cathode and the shielding electrode to the constant current source anode, forming a steep local electric field potential well in the cytochrome oxidase plaque region. At the same time, the arbitrary waveform generator outputs an asymmetric waveform with programmable rising edges and ramp and step characteristics. During the active charge recovery and cross-frame adaptive modulation stage, the hardware state machine controls the CMOS switch matrix to connect the active core electrode to the anode of the constant current source and the shielding electrode to the cathode of the constant current source. During the constant potential clamping and clearing phase, the hardware state machine controls the CMOS switch matrix to cut off the constant current drive circuit.

6. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 4, characterized in that, The central processing and control module controls the CMOS switch matrix to drive the designated active core electrode and shielding electrode for stimulation through a constructed routing table; wherein, the hardware routing table includes: electrode ID, physical coordinates, mask label, associated active electrode ID, number of shielding ring electrodes, switch status code and abnormal flag bit.

7. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 1, characterized in that, Methods for constructing routing tables include: The CMOS switch matrix sends bi-phase constant current detection pulses to the electrode array channel by channel in a time-division multiplexing manner, and performs residual voltage active clamping after the pulse to collect multi-dimensional physiological characteristics of each channel. Based on the aforementioned multi-dimensional physiological characteristics, a binary mask is generated using an adaptive dual-threshold algorithm to distinguish between cytochrome oxidase plaques and the interval regions between adjacent cytochrome oxidase plaques. The full-channel electrode initial screening is completed based on a binary mask to isolate non-functional area electrodes, while multiple independent CO Blob functional connectivity domains are separated by the 8-neighborhood connectivity rule. The number of shieldable electrodes is calculated within each CO Blob functional connectivity region, and active core electrodes and shielding electrodes are selected accordingly. Electrodes in non-connectivity regions are marked as isolation electrodes and set to a high-resistance state. A hardware routing table is constructed based on the active nuclear electrode, shielding electrode, and isolation electrode.

8. The three-phase clamped waveform based visual cortex color vision function reconstruction system according to claim 1, characterized in that, The calculation of the number of shieldable electrodes within each CO Blob functional connectivity region, and the selection of active core electrodes and shielding electrodes accordingly, includes: For each electrode within each connected domain, calculate its shielding resource count; wherein, the shielding resource count is defined as the number of other electrodes belonging to the same connected domain in the neighborhood within a preset radius centered on that electrode; Iterate through the number of shielding resources of all electrodes in each connected component and select the electrode with the largest number of shielding resources as the active core electrode of that connected component. All electrodes belonging to the same connected domain in the neighborhood of the active nuclear electrode are marked as shielding electrodes, thereby forming a shielding ring.

9. A method for visual cortex color vision function reconstruction based on three-phase clamped waveform, characterized in that, The method, applied to the visual cortex color vision function reconstruction system based on a three-phase clamping waveform as described in any one of claims 1 to 8, comprises: Acquire visual images and residual polarization voltage detected by the residual voltage monitor; Based on the acquired visual images and the residual polarization voltage detected by the residual voltage monitor, a color vision control command is generated, which controls the CMOS switch matrix to execute a three-phase closed-loop stimulation waveform on the stimulation load to realize the reconstruction of color vision function in the visual cortex; wherein, the three-phase closed-loop stimulation waveform is an asymmetric waveform envelope.

10. An electronic terminal, characterized in that include: One or more memories and one or more processors; The one or more memories are used to store computer programs; The one or more processors are connected to the memory and are used to run the computer program to perform the method as described in any one of claims 1 to 8.

Images