An all-implanted micro-brain control device and method

Abstract

The application discloses a kind of whole implantation micro brain control devices, the present application receives the model parameter of outside training completion using wireless communication device, and using micro processing device, the machine learning model of having trained completion is deployed in the FPGA computing device that has been implanted in, to realize the FPGA computing device that can complete nerve signal decoding, when the FPGA computing device receives nerve signal, target control signal can be directly calculated, and target control signal is sent to external equipment through microprocessor device and wireless communication device, to realize in vivo directly converting electroencephalogram nerve signal into target control signal and transmitting, increase system integration and avoid the power consumption of transmission original signal.The application also discloses a kind of whole implantation micro brain control method.

Description

Technical Field

[0001] This invention belongs to the field of brain-computer interface technology, specifically relating to a fully implantable micro brain control device and a fully implantable micro brain control method. Background Technology

[0002] Accurate, reliable, and stable acquisition of biological neural signals is an important prerequisite for conducting research on the behavior of organisms. Research on these biological neural signals can be used to analyze the intention signals of subjects or diagnose lesions of the nervous system, and has important clinical application value.

[0003] The brain is the most complex and active organ in the human body. Various neurological diseases, strokes, and injuries can damage brain tissue, causing impairments in motor and language functions, and affecting patients' communication and control with the outside world. Collecting patients' electroencephalogram (EEG) signals and analyzing them using artificial intelligence models holds promise for enabling direct interaction with the outside world through EEG signals, representing a new approach to functional reconstruction and disease treatment. Therefore, designing a system that monitors brain dynamics in real time and analyzes and controls brain activity accordingly is essential.

[0004] A neural signal acquisition device typically consists of acquisition electrodes, an analog front-end amplifier, an analog-to-digital converter, a microcontroller, and a data transmission device. Currently, more mature neural signal acquisition devices generally include implantable microelectrodes, an analog front-end amplifier, an analog-to-digital converter, a signal processing host, and a computer. This approach has advantages such as stable signal recording and strong processing capabilities, but it is generally larger in size and consumes more power.

[0005] Brain-computer interface (BCI) technology has shown significant advantages and potential in motor function reconstruction and the treatment of neurological or psychiatric diseases, and has now entered the clinical trial and application stage. Clinical applications require that the BCI system be fully implantable within the human body. This means that a microelectrode array is connected to a miniaturized neural signal acquisition and processing device. By fully implanting the system, neural signals are transmitted wirelessly, avoiding the risks of infection associated with percutaneous interfaces. This represents the trend in the development of integrated BCI devices.

[0006] Field-programmable gate arrays (FPGAs) play a crucial role in brain-computer interfaces, offering high flexibility and customizability, which are essential for processing complex neural signals and enabling real-time feedback.

[0007] Chinese patent application CN109222955A discloses an FPGA-based implantable neural signal acquisition device belonging to the fields of biosignal processing and analog electronics. This device mainly consists of implantable microelectrodes, a signal acquisition board, an acceleration and attitude sensor module, a digital potentiometer, a digital-to-analog converter module, an analog-to-digital converter module, a DIP switch, an FPGA control module, an Ethernet interface chip, an RJ45 interface, an SFP+ optical module, a USB control chip, SDRAM memory, FLASH memory, and external devices. This patent utilizes FPGA multi-boot reconfiguration technology to optimize the acquisition performance of brain neural signals; however, the acquisition device disclosed in this patent cannot directly obtain control commands.

[0008] Therefore, there is a need to design a miniature brain-computer interface system that can be fully implanted in the body, which can convert the collected brain signals into control commands and directly control external devices. Summary of the Invention

[0009] This invention provides a fully implantable miniature brain control device that can directly acquire brain neural signals and analyze them into target control signals to control external devices.

[0010] A specific embodiment of the present invention provides a fully implantable micro brain control device, including a microelectrode array and a packaging device. The packaging device includes a neural signal acquisition device, an FPGA computing device, a microprocessor device, a wireless communication device, a wireless charging receiver, and a battery integrated within the packaging shell.

[0011] The microelectrode array is electrically connected to the neural signal acquisition device and is used to transmit the detected brain neural electrical signals to the neural signal acquisition device.

[0012] The neural signal acquisition device is connected to the FPGA computing device and is used to receive brain neural electrical signals based on working instructions, digitally convert the brain neural electrical signals to obtain neural signals, and send the neural signals to the FPGA computing device.

[0013] The FPGA computing device is connected to the microprocessor device and is used to receive and store working instructions, send the stored working instructions to the neural signal acquisition device, deploy a machine learning model based on the trained model parameters, decode the neural signals into target control signals based on the machine learning model, and send the target control signals to the microprocessor device.

[0014] The microprocessor device is connected to the wireless communication device and is used to transmit working instructions and trained model parameters to the FPGA computing device. It is also used to parse the target control signal and send it to the wireless communication device.

[0015] The wireless communication device is wirelessly connected to an external device and is used to send the working instructions received from the external device and the trained model parameters to the microprocessor device. It is also used to send the parsed target control signal to the external device.

[0016] The wireless charging receiver charges the battery, and the battery supplies power to the neural signal acquisition device, FPGA computing device, microprocessor device and wireless communication device through the wireless charging receiver.

[0017] Furthermore, the FPGA computing device includes a control instruction receiving and processing module, a control instruction register group, a neural data preprocessing module, and a machine learning model decoding and computing module;

[0018] The control instruction receiving and processing module is used to unpack the work instructions and store the unpacked work instructions in the control instruction register group.

[0019] The control instruction register group is used to send the unpacked working instructions to the neural signal acquisition device to control the neural signal acquisition device to acquire neural signals.

[0020] The neural data preprocessing module is used to acquire raw neural signals based on the unpacked working instructions, and to preprocess the neural signals using different threshold comparison strategies to obtain neural spike signals.

[0021] The machine learning model decoding and calculation module is used to decode the target control signal from the neural spike signal based on the deployed machine learning model, and send the target control signal to the microprocessor device through the communication interface.

[0022] Furthermore, the microprocessor device includes a data transmission module, which includes a data receiving submodule and a data sending submodule;

[0023] The data receiving submodule includes a data receiving unit, a data parsing and pass-through module, and a target control signal channel. The data receiving unit is used to send the received target control signal to the data parsing and pass-through module. The data parsing and pass-through module is used to parse the target control signal and send the parsed target control signal to the target control signal channel. The parsed target control signal is then sent to the wireless communication device through the target control signal channel.

[0024] The data transmission submodule includes a command sending channel and a data transmission unit. The command sending channel is used to transmit the received work instructions and trained model parameters to the data transmission unit, and the data transmission unit sends the work instructions and trained model parameters to the FPGA computing device.

[0025] Furthermore, the microprocessor device also includes an FPGA firmware upgrade module and a wireless charging control module;

[0026] The FPGA firmware upgrade module is connected to the FPGA computing device and the wireless communication device respectively. It is used to receive the FPGA firmware upgrade file and upgrade command from the wireless communication device, store the FPGA firmware upgrade file and upgrade command, and upgrade the FPGA computing device by burning the upgrade command and upgrade file.

[0027] The wireless charging control module is connected to both the wireless charging receiver and the wireless communication device. It is used to transmit battery power information and charging status information sent by the wireless charging receiver to an external device via the wireless communication device based on the wireless charging control protocol. At the same time, it receives power-on / off control commands from the external device via the wireless communication device and sends the power-on / off control commands to the wireless charging receiver based on the wireless charging control protocol to control the wireless charging receiver to charge the battery.

[0028] Furthermore, the microprocessor device also includes a temperature acquisition module, which includes a temperature filtering, conversion and storage unit and a temperature sampling unit.

[0029] The temperature sampling unit is connected to the temperature filtering and conversion storage unit and the temperature sensor respectively. The temperature sampling unit is used to receive the temperature inside the package detected by the temperature sensor and send the detected temperature to the temperature filtering and conversion storage unit.

[0030] The temperature filtering, conversion, and storage unit is connected to a wireless communication device. It is used to filter and convert the detected temperature to obtain temperature data, and then send the temperature data to an external device via the wireless communication device to achieve temperature uploading and alarm.

[0031] Furthermore, the wireless charging receiver includes a receiving coil, a receiving matching + rectifier circuit, and a voltage regulator circuit. The receiving matching + rectifier circuit is connected to the receiving coil and the voltage regulator circuit respectively to enable the receiving coil to operate at a specified resonant point, and is also used to send the transmission power to the voltage regulator circuit. The voltage regulator circuit is connected to the battery to supply power to the battery and to supply power to the neural signal acquisition device, FPGA computing device, microprocessor device, and wireless communication device.

[0032] Furthermore, the wireless charging receiver is used to receive the transmission power sent from the wireless charging transmitter. The wireless charging transmitter includes an MCU, a DDS circuit, a power amplifier + resonant circuit, a transmitting coil, a transmitting coil voltage and current feedback signal processing circuit, and a voltage output adjustable DC-DC circuit.

[0033] The MCU controls the DDS circuit to output a sinusoidal signal of a specified frequency. The DDS circuit drives the power amplifier and resonant circuit through the output sinusoidal signal, so that the transmitting coil works at the resonant point to output transmitting power.

[0034] The transmitting coil voltage and current feedback signal processing circuit is connected to the power amplifier + resonant circuit and the MCU respectively, and is used to send the operating status signal of the transmitting coil to the MCU. The operating status signal includes whether the transmitting coil is working at the resonant point and the load status of the wireless charging receiver. The MCU modulates the DDS circuit and the voltage adjustable output DCDC circuit according to the operating status signal to regulate the transmitting power of the transmitting coil.

[0035] Furthermore, the encapsulation housing also integrates a high-density feedthrough, which is electrically connected to the microelectrode array via wires. The high-density feedthrough is connected to the neural signal acquisition device and is used to send neural electrical signals to the neural signal acquisition device.

[0036] The high-density feedthrough includes a feedthrough substrate and a flange, the feedthrough substrate and the flange are connected by brazing, the feedthrough substrate includes a porous ceramic substrate and a feedthrough electrode, the feedthrough electrode is deposited in the pores of the ceramic substrate.

[0037] Furthermore, the neural signal acquisition device, FPGA computing device, microprocessor device and battery are isolated from the wireless communication device and wireless charging receiver by a magnetic shielding sheet inside the package.

[0038] The top of the encapsulation shell is made of ceramic material, while the sidewalls and bottom are made of titanium metal. The ceramic material is connected to the sidewalls by brazing.

[0039] On the other hand, the present invention also provides a fully implantable microbrain control method, employing the aforementioned fully implantable microbrain control device, comprising:

[0040] The microelectrode array is implanted into the cerebral cortex, and the encapsulation device is implanted into the skull under the cerebral cortex.

[0041] The neural signal acquisition device receives working instructions from an external device and acquires neural signals based on the working instructions. It then transmits the acquired neural signals to the external device via a wireless communication device. On the external device, the received neural signals are used to train model parameters using machine learning methods to fit the corresponding behavioral signals until convergence is achieved and the trained model parameters are obtained. The behavioral signals include the movements of a certain part of the body or higher cognitive functions.

[0042] The trained model parameters are sent to the FPGA computing device via a wireless communication device and a microprocessor device, thereby deploying the machine learning model on the FPGA computing device.

[0043] New working instructions from external devices are sent to a neural signal acquisition device via a wireless communication device and a microprocessor device. The neural signal acquisition device acquires neural signals again based on the new working instructions and sends the neural signals to an FPGA computing device. The FPGA computing device decodes the received neural signals using a machine learning model to obtain target control signals, which include computer cursor control signals, robotic arm control signals, or speech synthesis control signals. The target control signals are then sent to the microprocessor device.

[0044] The target control signal data is parsed by the microprocessor and then transmitted to the wireless communication device. The parsed target control signal is then sent to an external device via the wireless communication device.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] This invention utilizes a microprocessor device to receive model parameters trained externally, and based on these parameters, deploys a machine model within an implanted FPGA computing device. This enables the implantation of an FPGA computing device capable of decoding neural signals. When the FPGA computing device receives neural signals, it can directly convert them into target control signals and transmit these signals to external devices via the microprocessor device and wireless communication device. This allows for the direct conversion and transmission of EEG neural signals into target control signals within the body, increasing system integration and avoiding the power consumption associated with transmitting raw neural signals. Attached Figure Description

[0047] Figure 1 Exploded view and 3D view of the miniature brain control device provided for specific embodiments of the present invention;

[0048] Figure 2 Exploded view and external structural diagram of the high-density feedthrough device provided for specific embodiments of the present invention;

[0049] Figure 3 A structural block diagram of the packaging device provided in a specific embodiment of the present invention;

[0050] Figure 4 A block diagram of an FPGA computing device provided for a specific embodiment of the present invention;

[0051] Figure 5 A block diagram of a microprocessor device provided for a specific embodiment of the present invention;

[0052] Figure 6 Block diagram of a wireless charging receiver and a wireless charging transmitter provided in a specific embodiment of the present invention;

[0053] Figure 7 A diagram illustrating the use scenario of the fully implantable micro brain control device provided in a specific embodiment of the present invention;

[0054] Figure 8 A flowchart illustrating the steps of a fully implantable micro-brain control method provided in a specific embodiment of the present invention.

[0055] Among them, there is a microelectrode array 1, a wire 2, a high-density feedthrough 3, a side wall of the package housing 4, a top of the package housing 5, an integrated device for neural signal acquisition, FPGA computing and microprocessor devices 6, a wireless communication device 7, a wireless charging receiver 8, a battery 9, a magnetic shielding sheet 10, a ceramic substrate 3-1, a feedthrough electrode 3-2, and a flange 3-3. Detailed Implementation

[0056] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be noted that the embodiments described below are intended to facilitate the understanding of the present invention and do not constitute any limitation thereof.

[0057] To directly convert brainwave signals into target control signals, a specific embodiment of this invention directly implants a decoder-capable FPGA computing device into the body and integrates it with a neural signal acquisition device, an FPGA computing device, a microprocessor device, a wireless communication device, a wireless charging receiver, and a battery. Figure 1 As shown in the figure, a fully implantable micro brain control device provided by a specific embodiment of the present invention includes a microelectrode array 1 and a packaging device. The packaging device includes a high-density power supply 3, a neural signal acquisition device, an integrated device 6 of an FPGA computing device and a microprocessor device, a wireless communication device 7, a wireless charging receiver 8 and a battery 9, all integrated within the packaging shell.

[0058] In a specific embodiment of the present invention, the microelectrode array 1 is connected to the high-density power supply 3 integrated in the packaging device via a wire 2. The wire 2 and the high-density power supply 3 are electrically connected by means of pressing or welding. The high-density power supply 3 is electrically connected to the neural signal acquisition device. The microelectrode array 1 transmits the acquired brain neural electrical signals to the neural signal acquisition device through the high-density power supply 3.

[0059] In one specific embodiment, the microelectrode array 1 provided in this embodiment is a microfilament electrode array, a silicon-based electrode array, or an electrode based on a flexible polymer. The electrode based on a flexible polymer is a combination of a silicon-based microelectrode array and a flexible ribbon cable based on a flexible polymer, realizing the floating implantation of the microelectrode array.

[0060] like Figure 2As shown, the high-density feed 3 provided in a specific embodiment of the present invention includes a feedthrough substrate and a flange 3-3. The feedthrough substrate and the flange are connected by brazing. The feedthrough substrate provided in this embodiment includes a ceramic substrate 3-1 and a feedthrough electrode 3-2, which is deposited in the pores of the ceramic substrate. In one embodiment, the ceramic substrate is composed of aluminum oxide or zirconium oxide.

[0061] A specific embodiment of the present invention provides a method for preparing a high-density feed 3, comprising:

[0062] Multiple ceramic sheet layers are obtained, and integral injection molding is performed by drilling holes in the ceramic sheet layers or by pre-inserting needles in the ceramic sheet powder in a mold. The diameter of the through holes in the ceramic sheet layers is between 0.05mm and 0.5mm, and the spacing between the through holes is between 0.1mm and 1.0mm. Multiple ceramic sheet layers with through holes are stacked and fired to obtain a ceramic substrate 3-1.

[0063] The first layer of conductive paste fills each through-hole, and the layers are stacked sequentially to fill the entire through-hole. Then, the Nth layer of conductive paste covers the outermost layer of the ceramic sheet. The sheet is then co-fired at a temperature of 1450℃ to 1600℃ to form the feed electrode 3-2, thus obtaining the feed substrate. The Nth layer of conductive paste, in its molten state, can penetrate into the gap between the ceramic substrate and the feed electrode, improving the airtightness between the ceramic substrate and the feed electrode.

[0064] The feedthrough substrate, co-fired with ceramic and conductive paste, is placed in the flange. A biocompatible composite metal coating is deposited on the surface of the feedthrough substrate in contact with the flange. The feedthrough substrate and the flange are mechanically connected by brazing using materials such as gold, titanium, and their alloys. The finished high-density feedthrough can achieve a sealing rating of less than 5×10-9 ATM.CC / SEC. Under the testing standards MIL-STD-202G and METHOD 301, the withstand voltage can reach 125V; under the testing standards MIL-STD-202 and METHOD 302, and the testing conditions of 100V DC, the insulation strength can reach 1GΩ, and it can withstand thermal shock testing from -65 to 200℃ without damage. The thin step extending outward from the upper end of the flange is designed with a thickness of 0.1 to 0.35mm, which allows for subsequent laser welding to implant shells of different thicknesses.

[0065] In a specific embodiment of the present invention, a neural signal acquisition device, an FPGA computing device, a microprocessor device, a wireless communication device 7, a wireless charging receiver 8, a magnetic shielding sheet 10, and a battery 9 are welded and assembled onto a plastic support of a packaged housing.

[0066] In this specific embodiment of the invention, the top and bottom of the encapsulation shell are made of ceramic material, i.e., a ceramic shell, and the sidewalls are titanium rings, i.e., a titanium shell. The wireless communication device 7 and the wireless charging receiver 8 are placed between the ceramic shell and the magnetic shielding sheet 10. A high-density feedthrough is passed through the countersunk hole in the titanium shell and soldered to the neural signal acquisition device, the FPGA computing device, and the microprocessor device, and then sealed with the ceramic shell. This specific embodiment of the invention employs a combined ceramic and titanium shell encapsulation technology, significantly increasing the efficiency of wireless communication and wireless charging.

[0067] Specific embodiments of the present invention provide a sealing step for the encapsulation housing, including:

[0068] Welding fixtures were placed in a laser welding machine to perform laser welding connections on the weld seams between the ceramic and titanium shells, and between the thin-walled step of the high-density feedthrough flange and the countersunk hole in the titanium shell. After welding, the sealing level of the fully implantable micro-brain control device can reach 5×10. -9 Below ATM.CC / SEC. The microelectrode array with leads is soldered and assembled with a high-density feedthrough. Using a tooling mold, the countersunk hole in the titanium shell, between lead 2 and the high-density feedthrough 3, is injection molded. The injection molding material can be biocompatible materials such as medical-grade silicone or medical-grade epoxy resin.

[0069] like Figure 3 As shown, the packaging device provided by the present invention includes a neural signal acquisition device, an FPGA computing device, a microprocessor device, a wireless communication device, a wireless charging receiver, and a battery integrated within the packaging housing.

[0070] In one specific embodiment, the wireless communication module adopts a low-power Bluetooth solution, the microprocessor adopts the Nordic nRF52840 chip, the FPGA adopts the Lattice ICE40 series chip, and the neural signal acquisition circuit adopts the Intan RHD2000 series chip.

[0071] The neural signal acquisition device provided in the specific embodiment of the present invention is connected to an FPGA computing device, and is used to start reading brain neural electrical signals based on the working instructions sent by the FPGA computing device, and to digitally convert the brain neural electrical signals to obtain neural signals, and at the same time send the neural signals to the FPGA computing device.

[0072] The FPGA computing device provided in this specific embodiment of the invention communicates with a microprocessor device via an SPI interface. The FPGA computing device receives work instructions sent from an external device from the microprocessor device and stores these instructions. Based on the stored work instructions, the FPGA computing device controls a neural signal acquisition device to begin reading neural signals. The FPGA computing device also receives trained model parameters sent from the external device from the microprocessor device, deploys a specified machine learning model, decodes the received neural signals into target control signals based on the machine learning model, and sends the target control signals back to the microprocessor device. This achieves direct conversion of neural signals into target control signals within the body, avoiding the power consumption associated with transmitting neural signals externally for decoding in existing technologies.

[0073] like Figure 4 As shown, the FPGA computing device provided in a specific embodiment of the present invention includes a control instruction receiving and processing module, a control instruction register group, a neural data preprocessing module, and a machine learning model decoding and computing module;

[0074] In this specific embodiment of the invention, the control instruction receiving and processing module is used to unpack the work instructions and store the unpacked work instructions in the control instruction register group.

[0075] The control instruction register group provided in the specific embodiments of the present invention is used to send working instructions to the neural signal acquisition device to control the neural signal acquisition device to acquire neural signals.

[0076] The neural data preprocessing module provided in the specific embodiments of the present invention is used to preprocess neural signals based on different threshold comparison strategies according to the unpacked working instructions to obtain neural spike signals.

[0077] The machine learning model decoding and calculation module provided in a specific embodiment of the present invention is used to decode the target control signal from the neural spike signal of the neural signal based on the deployed machine learning model, and send the target control signal to the microprocessor device through the communication interface. In one specific embodiment, the machine learning model is LSTM, and the machine learning model decoding and calculation module is an LSTM decoding and calculation module.

[0078] like Figure 5 As shown, the microprocessor device provided in a specific embodiment of the present invention includes a data transmission module, an FPGA firmware upgrade module, a wireless charging control module, a temperature acquisition module, and a microprocessor upgrade module.

[0079] The data transmission module provided in a specific embodiment of the present invention includes a data receiving submodule and a data sending submodule;

[0080] The specific embodiments of the present invention provide a data receiving submodule consisting of a data receiving unit, a data parsing and transmission module, and a target control signal channel. The data receiving unit is used to send the received target control signal to the data parsing and transmission module. The data parsing and transmission module is used to parse the target control signal and send the parsed target control signal to the target control signal channel. The parsed target control signal is then sent to a wireless communication device through the target control signal channel. Finally, the parsed target control signal is sent to an external device through the wireless communication device, thereby enabling direct brain-controlled control of external devices.

[0081] The data transmission submodule provided in a specific embodiment of the present invention includes a command sending channel and a data transmission unit. The command sending channel is used to transmit the received work instructions and trained model parameters to the data transmission unit, and the data transmission unit sends the work instructions and trained model parameters to the FPGA computing device.

[0082] The FPGA firmware upgrade module provided in the specific embodiments of the present invention includes an FPGA firmware upgrade command receiving channel, an FPGA firmware upgrade file receiving channel, a firmware upgrade file storage module, and an FPGA program burning timing execution unit.

[0083] The FPGA firmware upgrade command receiving channel and the FPGA firmware upgrade file receiving channel are respectively used to receive FPGA firmware upgrade files and upgrade commands sent from external devices via a wireless communication device. The firmware upgrade file storage module is used to store FPGA firmware upgrade files and upgrade commands. The FPGA program burning timing execution unit is electrically connected to the firmware upgrade file storage module and the FPGA computing device, respectively, and is used to burn the upgrade commands and upgrade files to the FPGA computing device via SPI communication.

[0084] The wireless charging control module provided in a specific embodiment of the present invention includes a battery power acquisition channel, a charging status channel, a power on / off control channel, a wireless charging control protocol, and a data receiving and transmitting unit. The microprocessor is equipped with the wireless charging control protocol and communicates with the wireless charging device via IIC communication to achieve data exchange, enabling battery information uploading and power on / off control of the fully implantable micro-brain control device.

[0085] The data receiving and transmitting unit receives battery power information and charging status information from the wireless charging receiving device, and transmits the battery power information and charging status information to the wireless communication device through the battery power acquisition channel and the charging status channel respectively based on the wireless charging control protocol. The battery power information and charging status information are then transmitted to an external device outside the body through the wireless communication device.

[0086] In a specific embodiment of the present invention, a power-on / off control command is received from an external device via a power-on / off control channel, and the power-on / off control command is sent to a data receiving and transmitting unit based on a wireless charging control protocol. The data receiving and transmitting unit then sends the power-on / off control command to a wireless charging receiving device to control the wireless charging receiving device to charge the battery.

[0087] The temperature acquisition module provided in a specific embodiment of the present invention includes a temperature status channel, a temperature filtering, conversion and storage unit and a temperature sampling unit. The temperature status channel is used to receive temperature data in real time and send the temperature data to a wireless communication device.

[0088] In a specific embodiment of the present invention, the temperature sampling unit is connected to the temperature filtering conversion and storage unit and the temperature sensor respectively. The temperature sampling unit is used to receive the temperature inside the packaged shell detected by the temperature sensor and send the detected temperature to the temperature filtering conversion and storage unit.

[0089] The temperature filtering, conversion, and storage unit provided in a specific embodiment of the present invention is connected to a wireless communication device. It is used to filter and convert the detected temperature to obtain temperature data, and then send the temperature data to an external device through the wireless communication device to realize temperature uploading and alarm.

[0090] The microprocessor upgrade module provided in a specific embodiment of the present invention includes a microprocessor upgrade channel and a microprocessor upgrade execution module. The microprocessor upgrade channel is used to receive upgrade instructions from external devices and send the upgrade instructions to the microprocessor upgrade execution module, so as to upgrade the microprocessor device through the microprocessor upgrade execution module.

[0091] like Figure 6 As shown in the figure, the wireless power supply device provided in a specific embodiment of the present invention includes a wireless charging receiver and a wireless charging transmitter. The wireless charging receiver is located inside the body, and the wireless charging transmitter is located outside the body.

[0092] The wireless charging receiver provided in a specific embodiment of the present invention includes a receiving coil, a receiving matching + rectifier circuit, and a voltage regulator circuit. The receiving matching + rectifier circuit is connected to the receiving coil and the voltage regulator circuit respectively to enable the receiving coil to operate at a specified resonant point and to send the transmission power to the voltage regulator circuit. The voltage regulator circuit is connected to a battery to supply power to the battery and to supply power to a neural signal acquisition device, an FPGA computing device, a microprocessor device, and a wireless communication device. In one specific embodiment, the voltage regulator circuit outputs a 5V±20% voltage to supply power to subsequent circuits.

[0093] The wireless charging receiver provided in this specific embodiment of the invention is used to receive transmission power sent from a wireless charging transmitter. The wireless charging transmitter consists of an MCU, a DDS circuit, a power + resonant circuit, a transmitting coil, a transmitting coil voltage and current feedback signal processing circuit, and a voltage-adjustable DC-DC converter circuit. The MCU controls the DDS circuit to output a sinusoidal signal of a specified frequency via a serial port, and can also control the amplitude of the output sinusoidal signal. The sinusoidal signal output by the DDS drives the power + resonant circuit to make the transmitting coil work at the resonant point, allowing the transmitting coil to output the maximum transmission power in the current state. The transmitting coil voltage and current feedback signal processing circuit provides feedback on the operating state of the transmitting coil, indicating whether the coil is working at the resonant point, and also provides feedback on the load state of the receiver. Based on the feedback signals, the MCU adjusts the DDS circuit and the voltage-adjustable DC-DC converter circuit in real time, enabling the overall circuit to dynamically respond to changes in the receiver current, resulting in a more suitable transmission power for the transmitter, thereby reducing overheating of the receiver circuit due to excessive transmission power.

[0094] like Figure 7 As shown, a specific embodiment of the present invention provides a usage scenario and a method for implanting a fully implantable micro brain-control device into the body, including:

[0095] The device is surgically implanted into the skull. The surgical procedure is as follows: The scalp at the implantation site is incised to expose the skull. A groove, roughly the same size as the fully implantable microbrain control device and leads, is ground at an appropriate location to hold the device and leads. The device is then fixed in place using skull screws. At the intended implantation location for the microelectrode, the skull and dura mater are sequentially cut, and the microelectrode is placed intracranially. After confirming the microelectrode placement, the dura mater is sutured, and the cut skull is backfilled. A secondary fixation is performed using a PEEK or titanium skull mesh and skull screws to secure the fully implantable microbrain control device, leads, and microelectrode. Finally, the scalp is sutured to complete the implantation surgery. An external charging device is used to wirelessly charge the fully implantable microbrain control device. The charging coil of the external charging device is placed against the scalp, roughly on top of the fully implantable microbrain control device. The distance between the charging coil and the device is between 1-5 cm. Using an external control device, wireless communication is established via Bluetooth with the fully implantable microbrain control device. Features are extracted from the neural electrical signals acquired by the fully implantable microbrain control device, and a machine learning model is applied to decode the features into target control signals. These target control signals are then transmitted externally via the microprocessor device and the wireless communication device.

[0096] like Figure 8 As shown, a specific embodiment of the present invention provides a fully implantable micro-brain control method, employing the aforementioned fully implantable micro-brain control device, comprising:

[0097] The microelectrode array is implanted into the cerebral cortex, and the encapsulation device is implanted into the skull beneath the cerebral cortex.

[0098] The neural signal acquisition device provided in the specific embodiment of the present invention receives working instructions from external devices and acquires neural signals based on the working instructions. When the neural signals reach a preset amount, the acquired neural signals are packaged into data packets and sent to the external devices through a wireless communication device.

[0099] On an external device, model parameters are trained using machine learning methods based on received neural signals to fit the corresponding behavioral signals until convergence is achieved, resulting in trained model parameters. The behavioral signals include movements of a part of the body, such as fine motor skills of the hand, or higher cognitive functions, such as language.

[0100] The trained model parameters are sent to the FPGA computing device via a wireless communication device and a microprocessor device, thereby deploying the machine learning model on the FPGA computing device.

[0101] New operating instructions from external devices are sent to the neural signal acquisition device via a wireless communication device and a microprocessor device. The neural signal acquisition device then acquires neural signals again based on the new operating instructions. When the acquired neural signals reach a preset amount, the acquired neural signals are sent to the FPGA computing device.

[0102] The FPGA computing device decodes the received neural signals using a machine learning model to obtain target control signals, which include computer cursor control signals, robotic arm control signals, or speech synthesis control signals, and sends the target control signals to the microprocessor device.

[0103] The target control signal data is parsed by the microprocessor and then transmitted to the wireless communication device. The parsed target control signal is then sent to an external device via the wireless communication device.

[0104] When the battery is low, the device of the present invention can be charged using a wireless charging device.

[0105] In one specific embodiment, the machine learning method provided by this invention includes linear and nonlinear classification or fitting algorithms.

[0106] The embodiments described above provide a detailed explanation of the technical solutions and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fully implantable micro-brain control device, characterized in that, It includes a microelectrode array and a packaging device, wherein the packaging device includes a neural signal acquisition device, an FPGA computing device, a microprocessor device, a wireless communication device, a wireless charging receiver, and a battery integrated within the packaging housing; The microelectrode array is electrically connected to the neural signal acquisition device and is used to transmit the detected brain neural electrical signals to the neural signal acquisition device. The neural signal acquisition device is connected to the FPGA computing device and is used to receive brain neural electrical signals based on working instructions, digitally convert the brain neural electrical signals to obtain neural signals, and send the neural signals to the FPGA computing device. The FPGA computing device is connected to the microprocessor device and is used to receive and store working instructions, send the stored working instructions to the neural signal acquisition device, deploy a machine learning model based on the trained model parameters, decode the neural signals into target control signals based on the machine learning model, and send the target control signals to the microprocessor device. The microprocessor device is connected to the wireless communication device and is used to transmit working instructions and trained model parameters to the FPGA computing device. It is also used to parse the target control signal and send it to the wireless communication device. The wireless communication device is wirelessly connected to an external device and is used to send the working instructions received from the external device and the trained model parameters to the microprocessor device. It is also used to send the parsed target control signal to the external device. The wireless charging receiver charges the battery, and the battery supplies power to the neural signal acquisition device, FPGA computing device, microprocessor device and wireless communication device respectively through the wireless charging receiver. The FPGA computing device includes a control instruction receiving and processing module, a control instruction register group, a neural data preprocessing module, and a machine learning model decoding and computing module. The control instruction receiving and processing module is used to unpack the work instructions and store the unpacked work instructions in the control instruction register group. The control instruction register group is used to send the unpacked working instructions to the neural signal acquisition device to control the neural signal acquisition device to acquire neural signals. The neural data preprocessing module is used to acquire raw neural signals based on the unpacked working instructions, and to preprocess the neural signals using different threshold comparison strategies to obtain neural spike signals. The machine learning model decoding and calculation module is used to decode the target control signal from the neural spike signal based on the deployed machine learning model, and send the target control signal to the microprocessor device through the communication interface. The microprocessor device includes a data transmission module, which includes a data receiving submodule and a data sending submodule. The data receiving submodule includes a data receiving unit, a data parsing and pass-through module, and a target control signal channel. The data receiving unit is used to send the received target control signal to the data parsing and pass-through module. The data parsing and pass-through module is used to parse the target control signal and send the parsed target control signal to the target control signal channel. The parsed target control signal is then sent to the wireless communication device through the target control signal channel. The data transmission submodule includes a command sending channel and a data transmission unit. The command sending channel is used to transmit the received work instructions and trained model parameters to the data transmission unit, and the data transmission unit sends the work instructions and trained model parameters to the FPGA computing device.

2. The fully implantable microbrain-controlled device according to claim 1, characterized in that, The microprocessor device also includes an FPGA firmware upgrade module and a wireless charging control module. The FPGA firmware upgrade module is connected to the FPGA computing device and the wireless communication device respectively. It is used to receive the FPGA firmware upgrade file and upgrade command from the wireless communication device, store the FPGA firmware upgrade file and upgrade command, and upgrade the FPGA computing device by burning the upgrade command and upgrade file. The wireless charging control module is connected to both the wireless charging receiver and the wireless communication device. It is used to transmit battery power information and charging status information sent by the wireless charging receiver to an external device via the wireless communication device based on the wireless charging control protocol. At the same time, it receives power-on / off control commands from the external device via the wireless communication device and sends the power-on / off control commands to the wireless charging receiver based on the wireless charging control protocol to control the wireless charging receiver to charge the battery.

3. The fully implantable microbrain-controlled device according to claim 1, characterized in that, The microprocessor device further includes a temperature acquisition module, which includes a temperature filtering, conversion and storage unit and a temperature sampling unit. The temperature sampling unit is connected to the temperature filtering and conversion storage unit and the temperature sensor respectively. The temperature sampling unit is used to receive the temperature inside the package detected by the temperature sensor and send the detected temperature to the temperature filtering and conversion storage unit. The temperature filtering, conversion, and storage unit is connected to a wireless communication device. It is used to filter and convert the detected temperature to obtain temperature data, and then send the temperature data to an external device via the wireless communication device to achieve temperature uploading and alarm.

4. The fully implantable microbrain-controlled device according to claim 1, characterized in that, The wireless charging receiver includes a receiving coil, a receiving matching + rectifier circuit, and a voltage regulator circuit. The receiving matching + rectifier circuit is connected to the receiving coil and the voltage regulator circuit respectively to enable the receiving coil to operate at a specified resonant point, and is also used to send the transmission power to the voltage regulator circuit. The voltage regulator circuit is connected to the battery to supply power to the battery and to supply power to the neural signal acquisition device, FPGA computing device, microprocessor device, and wireless communication device.

5. The fully implantable microbrain-controlled device according to claim 1, characterized in that, The wireless charging receiver is used to receive the transmission power sent from the wireless charging transmitter. The wireless charging transmitter includes an MCU, a DDS circuit, a power amplifier + resonant circuit, a transmitting coil, a transmitting coil voltage and current feedback signal processing circuit, and a voltage output adjustable DC-DC circuit. The MCU controls the DDS circuit to output a sinusoidal signal of a specified frequency. The DDS circuit drives the power amplifier and resonant circuit through the output sinusoidal signal, so that the transmitting coil works at the resonant point to output transmitting power. The transmitting coil voltage and current feedback signal processing circuit is connected to the power amplifier + resonant circuit and the MCU respectively, and is used to send the operating status signal of the transmitting coil to the MCU. The operating status signal includes whether the transmitting coil is working at the resonant point and the load status of the wireless charging receiver. The MCU modulates the DDS circuit and the voltage adjustable output DCDC circuit according to the operating status signal to regulate the transmitting power of the transmitting coil.

6. The fully implantable microbrain-controlled device according to claim 1, characterized in that, The encapsulation housing also integrates a high-density feedthrough, which is electrically connected to the microelectrode array via wires. The high-density feedthrough is connected to a neural signal acquisition device and is used to send neural electrical signals to the neural signal acquisition device. The high-density feedthrough includes a feedthrough substrate and a flange, the feedthrough substrate and the flange are connected by brazing, the feedthrough substrate includes a porous ceramic substrate and a feedthrough electrode, the feedthrough electrode is deposited in the pores of the ceramic substrate.

7. The fully implantable microbrain control device according to claim 1, characterized in that, The neural signal acquisition device, FPGA computing device, microprocessor device and battery are isolated from the wireless communication device and wireless charging receiver by a magnetic shielding sheet inside the package. The top of the encapsulation shell is made of ceramic material, while the sidewalls and bottom are made of titanium metal. The ceramic material is connected to the sidewalls by brazing.

Images