Multi-channel multi-mode active electrode integrated nerve probe chip of implantable brain-computer interface

Abstract

The invention discloses a multi-channel multi-mode active electrode integrated nerve probe chip for an implantable brain-computer interface, which comprises a non-implantable circuit part and an implantable electrode part, wherein the non-implanted circuit part integrates a multi-channel electroneurographic signal detection front end, a chemical signal detection front end, a multi-bit adjustable stimulation circuit, a reconfigurable switch array and digital control logic; the implantable electrode part is of a silicon needle structure, electrical electrodes and chemical electrodes are longitudinally distributed on the surface of the implantable electrode part, and dynamic gating and closed-loop regulation and control are supported. Through the electrode-circuit integrated design, the multi-mode function of synchronous high-precision detection and electrical stimulation output of the electroneurographic signals and the chemical signals is achieved, the problems that in the prior art, the mode is single, and the integration degree is low are solved, and the device has the advantages of being high in channel density and low in noise; 5 [mu] Vrms) and a fast closed-loop response (lt; the method is suitable for acute and long-term brain science research and nerve disease treatment.

Description

Technical Field

[0001] This invention relates to the fields of analog integrated circuit technology and biosensor microfabrication technology, specifically to a multi-channel, multimodal active electrode integrated neural probe chip that combines integrated circuit technology and microfabrication technology to record neural electrical signals, record neurochemical signals, and stimulate neural electrical stimulation for implantable brain-computer interfaces. Background Technology

[0002] The complexity and information processing capabilities of the human brain make the study of neurons and synapses crucial. It is estimated that the human brain contains approximately 80 billion neurons, 100 trillion synapses, and 100 different neurotransmitters. A deep understanding of its information encoding principles is essential for understanding and treating neurological diseases and developing neuromorphic computing methods. Therefore, large-scale simultaneous recording of neural electrical and chemical signals is key to revealing the mechanisms of brain behavior and activity.

[0003] Thanks to advancements in integrated circuit and biosensor microfabrication technologies, electrical and chemical electrodes can now be grown on existing circuitry, enabling the integration of electrodes and circuits and the development of highly integrated, multimodal electrode-circuit integrated chips. This multimodal integrated design not only allows for innovation at the system level, enhancing recording performance through electrode-circuit synergy, but also enables researchers to conduct in-depth studies of different types of neural signals within the same experimental framework, revealing more complex neural mechanisms. Furthermore, neural electrical stimulation modulation allows for specific stimulation of nerve cells as needed, enabling closed-loop system analysis of nerve cells in conjunction with neural recordings. Therefore, system-level neural probe chips integrating multiple functions have promising prospects for application in brain science research.

[0004] Currently, implantable electrodes can have thousands or even tens of thousands of channels. This technological advancement allows researchers to capture more comprehensive dynamic changes in the brain, especially in complex behavioral and cognitive tasks. While existing technologies have made progress in improving electrode recording density and spatiotemporal resolution, limitations remain in recording range and accuracy. First, most implantable electrodes are currently passive, requiring the recorded analog signals to be connected to additional circuit chips for digitization. Since the electrodes and circuits are separate, it's impossible to dynamically select recording electrodes using circuit technology to expand the recording range of the implanted electrodes. Therefore, the number of channels that can be recorded simultaneously is ultimately limited by the number of channels in the circuit chip. Furthermore, this separate electrode-circuit connection introduces significant parasitic interference into the signal recording channels, increasing crosstalk between channels and reducing the common-mode rejection ratio of the overall signal chain, making it difficult to achieve high-precision signal recording. In 2017, a research team at the Belgian Centre for Microelectronics Research (CMIC) released version 1.0 of the Neuropixels, an implantable silicon-based active electrode integrating electrodes and circuitry. This version features 960 recording electrode sites and 384 recording channels. It utilizes a programmable switch array for electrode selection, allowing for flexible adjustment of electrode positions within the recording channels according to experimental needs. This expands the recording range and improves the ability to monitor specific brain regions. Furthermore, the integrated design reduces parasitism and achieves higher recording precision compared to discrete systems. In 2021, the team developed version 2.0, a high-density silicon-based active electrode integrating 5120 recording electrode sites and 384 recording channels. The optimized recording circuitry further expands the recording range and improves recording precision, making it the highest-throughput implantable active electrode neural probe currently available. However, this active electrode can only record a single modality of neural electrical signals and cannot detect chemical signals such as neurotransmitters. This limitation in detection modality prevents researchers from comprehensively studying the interaction mechanisms of nerve cells and understanding the functional and pathological changes of the nervous system. Furthermore, the lack of integrated stimulation functions makes it unsuitable for closed-loop regulation and research of nerve cells. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a multi-channel, multimodal active electrode integrated neural probe chip for implantable brain-computer interfaces. This neural probe chip integrates electrical electrodes, chemical electrodes, reference electrodes, stimulation electrodes, electro- and chemical detection circuits, and stimulation circuits, possessing multimodal and closed-loop control functions, and further improving the system's integration. This invention is an electrode-circuit integrated system chip with built-in needle-like implantable electrodes, which can be directly implanted into animals, suitable for conducting acute or long-term animal experiments.

[0006] To achieve the above functions, the technical solution of the present invention is as follows:

[0007] An integrated neural probe chip with multi-channel, multimodal active electrodes for implantable brain-computer interfaces is divided into a non-implantable multi-channel circuit section and an implantable multi-channel needle-shaped electrode section.

[0008] The non-implantable circuit portion has an area of ​​less than 5mm*5mm and includes: a multi-channel neural electrical signal detection front-end, which consists of three cascaded voltage amplifiers and an analog-to-digital converter (ADC). The gain, bandwidth, and sampling rate of the voltage amplifiers and the ADC are all reconfigurable to meet different testing requirements. The multi-channel chemical signal detection front-end is divided into a potential method detection front-end and a cyclic voltammetry detection front-end. The potential method works by the interaction between the chemical electrode surface and ions in the biological fluid to form an electrical double layer, creating a potential difference between the electrode and the fluid. This can be used to detect ions (such as H+) in the biological fluid. + Na + K + Ca 2+ The potential method detection front-end structure consists of an inverter-based comparator and a time-to-digital converter (TDC). The cyclic voltammetry method measures current by applying voltage and can be used to detect chemical neurotransmitters (such as dopamine). Its detection front-end consists of a transimpedance amplifier, a voltage amplifier, and an ADC. The gain of the transimpedance amplifier is reconfigurable to meet detection requirements for different current ranges. In addition, it includes multi-bit adjustable voltage and current stimulation circuits for generating neural electrical stimulation; a bandgap reference circuit; a reconfigurable switch array circuit; a finite state machine; and digital control logic.

[0009] The implantable electrode portion consists of a single or multiple uniformly distributed silicon needles, with a needle width between 70 μm and 100 μm and a length of approximately 10 mm. Each silicon needle includes an electrode and an electrode selection control switch. Multiple electrical and chemical electrodes are arranged in two longitudinal rows on the surface of the silicon needle; multiple sets of reconfigurable reference / stimulation / current recording electrodes are evenly distributed on the surface of the silicon needle at certain intervals; in addition, the electrical electrodes can also be reconfigured as stimulation electrodes or reference electrodes to achieve recording, reference, or stimulation functions in different modes.

[0010] The electrical electrode is a square microelectrode, approximately 15μm x 15μm in size, used to acquire neural electrical signals; the chemical electrode is approximately 10μm x 20μm in size, and a corresponding sensitive membrane is deposited as needed to selectively detect ions in body fluids.

[0011] The integrated active electrode neural probe chip was designed using a 180nm SOI process. Subsequent microfabrication techniques were employed to grow the electrode and insulating layers, releasing the neural probes. The electrode materials are platinum, gold, titanium nitride, or iridium oxide, exhibiting good biocompatibility. The insulating layer is silicon nitride, ensuring excellent electrical insulation between the electrical electrodes while improving the interlayer capacitance of the chemical electrodes, making it easier to acquire chemical signals.

[0012] By combining reactive ion etching, deep reactive ion etching, and grinding thinning, the chip is separated and released from the entire silicon wafer to form an implantable neural probe.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0014] 1) By integrating electrical and chemical electrodes, it is possible to simultaneously detect neural electrical and chemical signals, providing researchers with more comprehensive neural signal data and helping to understand the functional and pathological changes of the nervous system more deeply.

[0015] 2) The integrated design of electrodes and circuits greatly improves the system integration, reduces chip area, lowers power consumption, and improves system stability and reliability.

[0016] 3) Dynamic selection of electrodes is achieved through a reconfigurable switch array circuit, expanding the recording range. Simultaneously, high-performance amplifiers and ADCs ensure high-precision signal recording. Through the reconfigurable switch array circuit and finite state machine, this chip can flexibly configure the operating states of each electrode and circuit module.

[0017] 4) The integrated stimulation circuitry allows for specific stimulation of nerve cells, and combined with neural recordings, enables closed-loop system analysis of nerve cells, providing a powerful tool for neuroscience research. Attached Figure Description

[0018] Figure 1 This is an overall block diagram of the multi-channel, multi-modal active electrode integrated neural probe chip according to an embodiment of the present invention;

[0019] Figure 2 This is a structural block diagram of the neural electrical signal recording channel according to an embodiment of the present invention;

[0020] Figure 3 This is a structural block diagram of the chemical signal potential method detection front end according to an embodiment of the present invention;

[0021] Figure 4 This is a structural block diagram of the chemical signal cyclic voltammetry detection front end according to an embodiment of the present invention;

[0022] Figure 5 This is a structural block diagram of the nerve electrical stimulation circuit according to an embodiment of the present invention;

[0023] Figure 6 This is a flowchart illustrating the subsequent micromachining steps and their functions in an embodiment of the present invention. Detailed Implementation

[0024] To more clearly illustrate the advantages and innovations of this invention, the technical solutions and implementation methods are further described below with reference to examples and accompanying drawings, but this should not limit the scope of protection of this invention.

[0025] This embodiment discloses an integrated neural probe chip with multi-channel, multimodal active electrodes for implantable brain-computer interfaces. This chip integrates electrical electrodes, chemical electrodes, reference electrodes, stimulation electrodes, and corresponding electro- and chemical detection and stimulation circuits, thereby achieving multimodal detection and closed-loop control functions and significantly improving the system's integration.

[0026] The following is an example of an active electrode neural probe chip with a single silicon needle, and the overall framework diagram is shown below. Figure 1 As shown, the multi-channel, multi-modal active electrode integrated neural probe chip includes a non-implantable circuit section (1) and an implantable electrode section (2). The non-implantable circuit section (1) is integrated into a chip with an area of ​​less than 5mm × 5mm and includes a neural electrical signal detection front-end (4), a chemical signal detection front-end (9), a neural electrical stimulation module (8), a power supply and bias circuit (12), a chip digital control module (13), a communication module (14), and a parallel-to-serial conversion module (15). The implantable electrode section (2) consists of one or more silicon needles with a width between 70μm and 100μm and a length of about 10mm. The surface of the needles has electrical electrodes, chemical electrodes, reference electrodes, and stimulation electrodes distributed longitudinally, and is connected to the non-implantable circuit section through an internal active switch (17).

[0027] like Figure 2 As shown, the neural electrical signal detection front end (4) consists of a three-stage reconfigurable voltage amplifier and an ADC, whose gain, bandwidth and ADC sampling rate can be reconfigured according to the test requirements.

[0028] First stage amplifier (18): The input impedance and equivalent input noise can be dynamically adjusted by the digital control module (13). The input impedance range is 1GΩ to 10GΩ, and the noise level is less than 5μVrms (0.1Hz-10kHz).

[0029] Second stage amplifier (19): The low pass frequency can be configured to 5kHz (action potential band) or 300Hz (local field potential band), with a bandwidth error of less than ±5%.

[0030] Third stage amplifier (20): The gain is adjustable from 20dB to 60dB, and the high-pass frequency is adjusted by a 4-bit control word (1Hz-300Hz);

[0031] ADC(21): Dynamically adjustable sampling rate (1kSPS to 50kSPS), 12-bit resolution, signal-to-noise ratio (SNR) greater than 70dB.

[0032] Each detection front end supports 4 electrode channels, for a total of 256 channels, which are output to the host computer (16) via the parallel-to-serial conversion module (15).

[0033] In this embodiment, there are a total of 64 groups, consisting of 4 voltage signal recording channels (5), 1 bias circuit (6), and 1 front-end circuit control module (7), thus there are a total of 256 neural voltage signal recording channels. The 256 voltage signal recording channels (5) are used to record and digitize neural electrical signals, which are output to the external host computer control and data processing module (16) through 8 parallel-to-serial conversion circuits (15); the bias circuit (6) provides voltage bias for the 4 voltage signal recording channels (5) in the group, and the front-end circuit control module (7) is used to configure the working state of the voltage signal recording channels (5).

[0034] Chemical signal detection front end (9) is divided into potential method detection front end ( Figure 3 ) and cyclic voltammetry detection front end ( Figure 4 These are used to detect ions and chemical neurotransmitters in biological fluids, respectively. In this embodiment, the chemical signal detection front end (9) includes 8 sets of potential method (10) detection circuits to detect chemical voltage signals and 1 set of cyclic voltammetry detection circuit (11) to detect chemical current signals.

[0035] The nerve electrical stimulation module (8) outputs multiple adjustable voltage or current stimulations, wherein the voltage stimulation range is -2.5V to 2.5V and the current stimulation range is 0 to 500μA. Figure 5 As shown, it includes:

[0036] Multi-bit current DAC (27): Generates 0-500μA programmable current with a step accuracy of 1μA;

[0037] Transimpedance amplifier (28): converts current to a voltage signal from -2.5V to 2.5V;

[0038] The gating switch (29) outputs the stimulation signal to any electrical electrode with a switching time of less than 1 μs. The stimulation parameters (amplitude, frequency, waveform) are set by the host computer (16), supporting single pulse, biphasic pulse and custom waveform.

[0039] The power supply and bias circuit (12) consists of a bandgap reference circuit and a multi-channel current generation circuit, which generates a fixed voltage bias and current output inside the chip to provide bias voltage and bias current for other modules inside the chip.

[0040] The chip digital control module (13), communication module (14), and host computer control and data processing module (16) work together. The host computer control and data processing module (16) sends control commands to the communication module (14). After receiving and converting the commands, the communication module (14) generates control words and sends them to the chip digital control module (13). The chip digital control module (13) generates corresponding status control codes to control the working status of each module. At the same time, the working status of each module can also be received by the host computer control and data processing module (16) through the communication module (14) so ​​that the chip user can judge whether the various configurations of the current chip are correct.

[0041] The implantable electrode part 2 consists of one or more silicon needles, with a width between 70 μm and 100 μm and a length of approximately 10 mm. Each silicon needle integrates an electrode and an electrode selection control switch, realizing a longitudinal two-row distribution of electrical and chemical electrodes on the surface of the silicon needle. These electrodes can be flexibly configured for recording, reference, or stimulation functions to meet different experimental needs. In this embodiment, there are 1024 electrical / chemical electrodes and active circuits (17) and 4 sets of configurable reference / stimulation / current recording electrodes (3). Every 4 electrical electrodes are connected to the same neural voltage signal recording channel (5), with 256 electrodes forming a group, for a total of 4 groups. The control word can be output by the chip digital control module (13) to control whether the active switch below the electrode is turned on, allowing arbitrary selection and connection to the corresponding neural voltage signal recording channel (5). At the same time, any electrical electrode can be configured to connect to the neural electrical stimulation module (8) through the chip digital control module (13) to realize the stimulation output function of any electrode site. 1024 chemical electrodes, in groups of 128, are connected to the same time-to-digital converter. The chemical voltage sites connected are controlled by a gating switch, and the corresponding chemical voltage signals are digitized and then output to the host computer control and data processing module (16) via the communication module (13).

[0042] In this embodiment, a total of 4 sets of configurable reference / stimulation / current recording electrodes (3) are evenly distributed on the probe and can be configured to different functions. First, they can function as reference electrodes when recording nerve voltage signals to provide voltage reference for the recording site. Second, they can be connected to the nerve electrical stimulation module (8) to output voltage or current stimulation to realize closed-loop nerve control and recording. Third, they can be used as recording electrodes for chemical current signals and input to the cyclic voltammetry detection circuit (11).

[0043] Figure 2This is a block diagram of the neural voltage signal recording channel (5) in this embodiment. The neural voltage recording channel consists of three cascaded operational amplifiers and a high-efficiency analog-to-digital converter. The first-stage operational amplifier (18) has adjustable power consumption and input impedance, which can be flexibly configured in different recording scenarios, and can change the input impedance and equivalent input noise of the overall recording channel; the second-stage operational amplifier (19) has adjustable low-pass frequency, which can be configured as the neural electrical signal frequency band (low-pass 5KHz) and the local field potential signal frequency band (low-pass 300Hz); the third-stage operational amplifier (20) has adjustable gain and high-pass frequency, which can adjust the amplification factor of the operational amplifier according to the actual recording, and realize the large value input of the analog-to-digital converter as much as possible without affecting the recording; the high-pass frequency is 4-bit adjustable, and the adjustable range is (1Hz-300Hz); the sampling rate of the high-efficiency analog-to-digital converter (21) is adjustable, and the sampling rate can be adjusted according to the actual recorded signal frequency band, and the recording power consumption during actual operation can be dynamically adjusted. Therefore, the overall neural voltage signal recording channel (5) can realize the recording of individual local field potential signals, individual action potential signals, and full-band neural electrical signals.

[0044] Figure 3 This is a block diagram of the specific structure of the potential method detection front end (10) in this example, showing ions in biological fluids (such as H+). + Na + K + Ca 2+ The signal is converted into a voltage signal by the chemical electrode, superimposed with the triangular wave input and input to the comparator (22), and converted into a PWM wave signal in the time domain. Then, it is converted into a digital signal by the time-to-digital converter (23) and output to the host computer control and data processing module (16).

[0045] Figure 4 This is a block diagram of the specific structure of the cyclic voltammetry detection front end (11) in this example. When the cyclic voltammetry is used, it is a positive and negative triangular wave voltage scan. Therefore, the transimpedance amplifier (24) in this example is powered by ±2.5V. The chemical current signal is converted into a voltage signal with a voltage range of -2.5V to 2.5V by the transimpedance amplifier (24). However, the power supply voltage of the high-efficiency analog-to-digital converter (26) is 1.8V, which can only realize the input of the voltage range of 0-1.8V. Therefore, it is necessary to convert the voltage range to 0-1.8V by the voltage amplifier and input it to the high-efficiency analog-to-digital converter for digitization.

[0046] Figure 5The diagram below shows the structure of the nerve electrical stimulation module (8) in this example. A multi-bit current digital-to-analog converter (27) generates a multi-bit adjustable stimulation current. The stimulation current is converted into a stimulation voltage by a transimpedance amplifier (28) and then output to the stimulation electrode via a gating switch (29). The magnitude and frequency of the stimulation current and voltage can be controlled by the chip digital control module (13).

[0047] In this embodiment, the integrated active electrode neural probe chip is designed using a 180nm SOI process and manufactured by a wafer foundry to obtain an 8-inch wafer. The wafer needs to be processed to release the neural probes, and then the neural probes are wire-bonded to form a system prototype that is easy to implant. Figure 6 The flowchart outlines the steps for releasing implantable neural probes in subsequent microfabrication, including electrode growth, passivation layer growth, probe etching, silicon wafer thinning, and deposition of an insulating layer.

[0048] While the above examples describe specific embodiments of the present invention, those skilled in the art should understand that these are merely illustrative examples, and various changes or modifications can be made to these embodiments without departing from the essence of the present invention. Therefore, the scope of protection of the present invention is defined by the appended claims.

Claims

1. A multi-channel, multimodal active electrode integrated neural probe chip for implantable brain-computer interfaces, comprising, The non-implantable circuit portion and the implantable electrode portion are characterized in that the non-implantable circuit portion includes: The multi-channel neural electrical signal detection front end consists of three cascaded voltage amplifiers and an analog-to-digital converter. The gain and bandwidth of the voltage amplifiers and the sampling rate of the ADC can be dynamically adjusted. A multi-channel chemical signal detection front end includes a potential method detection front end and a cyclic voltammetry detection front end. The potential method detection front end includes an inverter-based comparator and a time-to-digital converter (TDC) for converting the voltage signal of the chemical electrode into a digital signal. The cyclic voltammetry detection front end includes a transimpedance amplifier, a voltage amplifier, and an ADC for converting the chemical current signal into a digital signal. A multi-position adjustable voltage / current stimulation circuit is used to output nerve electrical stimulation signals; Reconfigurable switch array circuit for dynamically selecting electrodes; The bandgap reference circuit provides stable voltage and current bias for the chip; Finite state machines and digital control logic are used to coordinate the working states of each module; The implantable electrode portion comprises one or more silicon needles, with electrical and chemical electrodes longitudinally distributed on the surface of the silicon needles. The electrical and chemical electrodes are connected to the non-implantable circuit portion via the reconfigurable switch array circuit. The electrical electrodes are square microelectrodes used to record neural electrical signals. A sensitive membrane is deposited on the surface of the chemical electrodes to detect ions or chemical neurotransmitters in biological fluids. The non-implantable circuit section and the implantable electrode section are connected by an active switch integrated inside the silicon needle, realizing multimodal functions of electrical signal recording, chemical signal detection, and electrical stimulation output, and completing closed-loop regulation through the digital control logic.

2. The neural probe chip according to claim 1, characterized in that, The area of ​​the non-implantable circuit portion is less than 5mm × 5mm, and the width of the silicon needle in the implantable electrode portion is 70μm to 100μm, and the length is 10mm.

3. The neural probe chip according to claim 1, characterized in that, The electrical electrode has a size of 15μm × 15μm, and the chemical electrode has a size of 10μm × 20μm.

4. The neural probe chip according to claim 1, characterized in that, The potential method detection front end generates a PWM wave by superimposing a triangular wave input and a chemical electrode signal, and converts it into a digital signal using the TDC; the cyclic voltammetry detection front end converts the chemical current signal into a voltage signal through a transimpedance amplifier, adjusts it to the ADC input range through a voltage amplifier, and finally digitizes it using the ADC.

5. The neural probe chip according to claim 1, characterized in that, The multi-position adjustable voltage / current stimulation circuit includes: A multi-bit current digital-to-analog converter (DAC) is used to generate programmable stimulation currents; A transimpedance amplifier converts the stimulation current into a stimulation voltage; The selector switch outputs the stimulation signal to the target electrode.

6. The neural probe chip according to claim 5, characterized in that, The voltage output range of the multi-position adjustable voltage / current stimulation circuit is -2.5V to 2.5V, and the current output range is 0 to 500μA.

7. The neural probe chip according to claim 1, characterized in that, The implantable electrode portion also includes a reference electrode and a stimulation electrode, which are dynamically configured with respect to recording, reference, or stimulation functions via the reconfigurable switch array circuit.

Images