A head-mounted neural stimulation recording device for small animals and its synchronous control method

Abstract

This invention discloses a head-mounted neural stimulation recording device for small animals and a synchronous control method. The device includes a three-electrode interface, a constant current source powered by a single positive high voltage, a four-switch H-bridge, a local field potential acquisition module, a load voltage sampling module, and a microcontroller. After the first stimulation phase, the high and low side switches of the same bridge arm are turned on, while the other bridge arm is turned off, opening the stimulation load and creating a local current path for the constant current source that does not pass through the load. After the second stimulation phase, only the low side switch connected to the second electrode is turned on to bias the tissue common-mode potential. The microcontroller synchronously controls the H-bridge state, input blanking, and in-phase load voltage sampling using hardware timing events to reduce the saturation recovery spike of the constant current source and recover the local field potential acquisition within the allowable input common-mode range. This device is suitable for long-term neural stimulation and recording experiments in freely moving small animals.

Description

Technical Field

[0001] This invention relates to the fields of neural engineering, biomedical electronic devices, and experimental neuromodulation technologies, and particularly to a head-mounted neural stimulation recording device for small animal experiments such as mice and rats, and its synchronous control method. This device can be used for electrical stimulation, local field potential signal acquisition, impedance-related parameter detection, wireless data transmission, and stimulation parameter control under free-movement or low-constraint experimental conditions. Background Technology

[0002] Deep brain stimulation (DBS) is an important neuromodulation technique that has received widespread attention in the treatment of movement disorders and other diseases, as well as in neuroscience research. To investigate the relationship between stimulation parameters, neural activity, behavioral states, and electrode-tissue interfaces, animal experiments often require simultaneous stimulation, neural signal recording, and state monitoring over extended periods.

[0003] Local field potentials can reflect the collective neural activity of a target brain region and are often used to study neural rhythms and stimulus responses. Electrode or tissue impedance parameters can reflect electrode interfaces, tissue states, or slowly varying physiological states, and are valuable in slow-timescale experiments. Therefore, in small animal experiments, the ability to simultaneously perform biphasic electrical stimulation, local field potential recording, load voltage sampling, and wireless data transmission on the same head-mounted device will facilitate long-term neuromodulation experiments.

[0004] In existing systems, some devices can achieve wireless recording, while others can achieve wireless or programmable stimulation. Some systems also focus on closed-loop stimulation algorithms, multi-signal fusion, local field potential artifact suppression, or H-bridge neural electrical stimulation. However, for freely moving small animals, headgear quality, power supply time, stimulation compliance voltage, microvolt-level recording noise, stimulation artifact isolation, impedance correlation detection, and wireless links all need to be simultaneously constrained. If these functions are implemented by multiple devices, it can easily increase the animal's burden, introduce cable constraints, or cause time asynchrony between different signals.

[0005] For example, CN110167630A discloses a head-mounted closed-loop neurofeedback stimulation device including a stimulation circuit, a brain signal monitoring front-end, a microcontroller, and a communication unit; CN113713255B discloses a deep brain stimulation system combining local field potential sensing, stimulation, wireless communication, and impedance detection. These documents reflect the development status of head-mounted stimulation, local field potential recording, wireless communication, and impedance detection in related technological fields.

[0006] Regarding the specific coordination of stimulation and recording, US10307594B2 discloses a current-to-analog converter, an H-bridge, and states such as biphasic stimulation, phase interval, and discharge; CN121055764A discloses a four-switch H-bridge and low-side switch discharge after stimulation; US6690974B2 discloses input blanking and holding capacitance during stimulation; US9101767B2 discloses sampling load-related voltage and calculating impedance at a determined moment of stimulation phase; and US11931579B2 discloses coupling a reference voltage to tissue via selected electrodes to assist in the sensing of neural signals. These schemes each relate to some of the functions of this application, but the following coordinated state combination is not explicitly given simultaneously in the above-mentioned literature: turning on the high-side switch and the low-side switch of the same bridge arm and turning off the other bridge arm during the phase interval, while keeping the stimulation load open, to form a local current path for the constant current source that does not pass through the stimulation load; then reusing one of the low-side switches of the H-bridge for tissue common-mode bias during non-stimulation recording, and coordinating the H-bridge state, input blanking and load voltage sampling by the same hardware timing event sequence.

[0007] Biphasic stimulation can employ either positive and negative power rails, or a single positive high-voltage rail combined with an H-bridge to change the direction of the load current. In the latter configuration, if the stimulation load and the constant current source current path are simultaneously disconnected during the phase interval, while the constant current source remains at a non-zero setting, the output node of the constant current source may shift towards the compliance voltage limit due to parasitic capacitance charging, resulting in a saturation recovery spike in the next stimulation phase. Furthermore, if the tissue common-mode potential does not return to the allowable range at the recording front end after stimulation, reclosing the input blanking switch may still overload the preamplifier.

[0008] Therefore, there is a need for a neural stimulation recording device suitable for head-mounted installation in small animals and its synchronous control method, which can simultaneously satisfy the open circuit of the stimulation load and the controlled operation of the constant current source during the phase interval, establish tissue common mode bias during non-stimulation recording, and make the H-bridge switching, input blanking and load voltage sampling have a time sequence relationship with respect to the stimulation phase. Summary of the Invention

[0009] The purpose of this invention is to provide a small animal head-mounted neural stimulation recording device and its synchronous control method to solve the problems of the constant current source easily entering saturation during the phase interval in single positive high-voltage rail constant current stimulation, the uncertainty of the recorded input common mode potential after stimulation, and the difficulty in unifying the timing of stimulation, input blanking and load voltage sampling.

[0010] To achieve the above objectives, this invention provides a head-mounted neural stimulation recording device for small animals. The device includes an electrode interface, a stimulation output module, a neural signal acquisition module, a load voltage sampling module, an analog-to-digital converter, and a microcontroller. It may also include a wireless communication module and a power supply module. The stimulation output module includes a constant current source powered by a single positive high-voltage power supply and an H-bridge switching network. The two arms of the H-bridge switching network are respectively connected to the two ends of the stimulation electrode pair. Each arm includes a high-side switch connected to the output of the constant current source and a low-side switch connected to a reference node. The neural signal acquisition module includes an input blanking switch and a preamplifier. The load voltage sampling module connects the constant current source output to a detection node between the constant current source output and the H-bridge switching network. The microcontroller coordinates the control of the H-bridge switch state, input blanking, and detection node voltage sampling through an event sequence triggered by a hardware timer.

[0011] In one embodiment, the electrode interface includes a first electrode connection terminal, a second electrode connection terminal, and a third electrode connection terminal. The first electrode connection terminal and the second electrode connection terminal form a stimulation electrode pair, and the first electrode connection terminal and the third electrode connection terminal form a local field potential recording electrode pair.

[0012] In one embodiment, the H-bridge switching network includes four controlled switches. A first stimulation phase is formed by the conduction of the high-side switch of the first bridge arm and the low-side switch of the second bridge arm. A phase-interval state is formed by the simultaneous conduction of the high-side and low-side switches of the same first bridge arm and the disconnection of two switches in the second bridge arm. At this time, the second electrode connection is in an open-circuit state, and no stimulation current flows through the stimulation load; the constant current source output is connected to the reference node via the high-side and low-side switches of the first bridge arm, forming a local current path that does not pass through the stimulation load. A second stimulation phase is formed by the conduction of the low-side switch of the first bridge arm and the high-side switch of the second bridge arm. After the second stimulation phase, at least two low-side switches can be conducted to form a stimulation-end discharge state, and then only the low-side switch connected to the second electrode connection is kept conducting to form a non-stimulation recording bias state. In one specific switching sequence, the stimulation-end discharge state briefly conducts all four switches, with two low-side switches connecting the first and second electrode connections to the reference node; in another switching sequence, only two low-side switches are conducted. This state sequence keeps the constant current source in a controlled operating state during the phase interval and biases the tissue common-mode potential to the input range allowed by the neural signal acquisition module before resuming local field potential acquisition.

[0013] In one embodiment, an input blanking switch is located between the electrode interface and the DC servo preamplifier. During non-stimulation periods, the input blanking switch is turned on, allowing a local field potential signal to enter the DC servo preamplifier; during at least a portion of the stimulation window, the input blanking switch is turned off to isolate stimulation artifacts and protect the preamplifier.

[0014] In one embodiment, the microcontroller acquires load voltage samples at predetermined time points within the stimulation phase. The load voltage samples can be corrected by voltage divider ratio and reference node bias, and then combined with the stimulation current to obtain impedance-related parameters.

[0015] In one embodiment, the microcontroller acquires raw local field potential data at a high first sampling rate, marks, removes or reconstructs samples affected by input blanking, and then performs digital filtering and downsampling to form a wireless local field potential data stream with a lower sampling rate.

[0016] In one embodiment, the power module includes a lightweight battery, a low-voltage power supply path, and a boost circuit. The low-voltage power supply path powers the microcontroller, the wireless communication module, and the neural signal acquisition module, while the boost circuit provides a stimulus compliance voltage to the stimulus output module. The lightweight battery may include a series-connected zinc-air battery, and the device may be mounted on a flexible circuit board, a rigid-flex circuit board, or a small rigid circuit board.

[0017] This invention also provides a synchronous control method for a small animal head-mounted neural stimulation recording device. The method utilizes a hardware timer to trigger an H-bridge switch sequence sequentially including a first stimulation phase, a phase-interval state, a second stimulation phase, and a non-stimulation recording bias state. In the phase-interval state, the high-side and low-side switches of the same bridge arm are turned on, while the two switches of the other bridge arm are turned off, keeping the stimulation load open and allowing the constant current source to form a local current path that does not pass through the stimulation load. The method further includes disconnecting the input blanking switch within the stimulation window, acquiring the detection node voltage between the constant current source output and the H-bridge switch network at a predetermined time during the stimulation phase, closing the input blanking switch after entering the non-stimulation recording bias state, and processing the local field potential samples affected by the input blanking.

[0018] Compared with the prior art, the present invention has at least the following beneficial effects: First, in a phase-interval state, the high-side and low-side switches of the same bridge arm are simultaneously turned on, while the other bridge arm is turned off, keeping the stimulation load open while providing a defined local current path for the constant current source. Under the prototype test conditions of this application, no further stimulation phase recovery spikes that appeared in the original open-circuit state were observed after adopting this state.

[0019] Second, the first stimulation phase, phase-interval state, second stimulation phase, stimulation end discharge state, and non-stimulation recording bias state are organized into a continuous switching sequence, so that the same H-bridge can not only complete the biphasic stimulation and phase-interval constant current source state control, but also bias the tissue common mode potential through a single low-side switch before the local field potential acquisition is restored.

[0020] Third, by synchronously controlling the H-bridge switch state, input blanking switch, and detection node voltage sampling through the same hardware timing event sequence, the sampling time of the input blanking boundary and load voltage remains fixed relative to the stimulus phase, thus avoiding the timing-dependent scheduling of the aforementioned wireless communication tasks.

[0021] Fourth, by using a single positive high-voltage rail to generate biphasic stimulation, and by using the same analog-to-digital converter to collect local field potential and detection node voltage in a time-sharing manner, the high-voltage power rail and analog-to-digital conversion resources in the head-mounted device can be reduced compared to using positive and negative high-voltage rails and independent sampling devices. Attached Figure Description

[0022] Figure 1 This is a block diagram of the overall structure of a small animal head-mounted neural stimulation recording device according to an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of the three-electrode stimulation and recording connection according to an embodiment of the present invention.

[0024] Figure 3 This is a schematic diagram of a single positive high-voltage rail H-bridge stimulation output module and its switching state according to an embodiment of the present invention.

[0025] Figure 4 This is a schematic diagram of a local field potential acquisition module according to an embodiment of the present invention.

[0026] Figure 5 This is a timing diagram illustrating the first stimulus phase, phase interval state, second stimulus phase, stimulus termination short-circuit discharge state, non-stimulation recording bias state, input blanking, load voltage sampling, and local field potential data processing according to an embodiment of the present invention.

[0027] Figure 6 This is a schematic diagram of a power module and power supply path according to an embodiment of the present invention.

[0028] Figure 7 This is a schematic diagram of the functional partitioning and isolation arrangement of a head-mounted circuit board according to an embodiment of the present invention.

[0029] Explanation of reference numerals in the attached figures: 10 Electrode interface; E1 First electrode connection terminal; E2 Second electrode connection terminal; E3 Third electrode connection terminal; 20 Stimulation output module; 20a Constant current source; 20b H-bridge switching network; 30 Neural signal acquisition module; 31 Input blanking switch; 32 DC servo preamplifier; 33 Gain and filtering circuit; 34 Bias hold network; 40 Load voltage sampling module; 50 Analog-to-digital converter; 60 Microcontroller; 70 Wireless communication module; 80 Power supply module; 80a Battery pack; 80b Low-voltage power supply path; 80c Boost circuit; 81 Current setting unit; 90 Circuit board; 100 Headset; 200 External host; S1 First high-side switch; S2 First low-side switch; S3 Second low-side switch; S4 Second high-side switch; VLOAD Detection node; VREF Low-side reference node. Detailed Implementation

[0030] The following describes specific embodiments of the present invention in conjunction with the accompanying drawings. The device models and parameters described below are used to illustrate the connection relationships and working process of the prototype, and do not imply that every embodiment must use the same devices or values; the range of substitutions should be limited to the point that the corresponding structural relationships and control states can still be achieved after substitution.

[0031] like Figure 1 As shown, in one embodiment, the small animal head-mounted neural stimulation recording device 100 includes an electrode interface 10, a stimulation output module 20, a neural signal acquisition module 30, a load voltage sampling module 40, an analog-to-digital converter 50, a microcontroller 60, a wireless communication module 70, and a power supply module 80. Except for the removable battery and mechanical fasteners, all of the above modules are mounted on the same circuit board 90 suitable for small animal head-mounting. The circuit board 90 can be a flexible circuit board to reduce weight and facilitate head-mounting, or it can be a rigid-flexible hybrid circuit board or a small rigid circuit board.

[0032] Device 100 can communicate with an external host via wireless communication module 70. The external host can send stimulation amplitude, pulse width, stimulation frequency, blanking window, sampling rate, downsampling factor, or start / stop commands. Device 100 can send local field potential data, load voltage samples, impedance-related parameters, stimulation status, battery status, or diagnostic data to the external host.

[0033] like Figure 2 As shown, the electrode interface 10 includes a first electrode connection terminal E1, a second electrode connection terminal E2, and a third electrode connection terminal E3. E1 and E2 form a stimulation electrode pair for applying electrical stimulation to a target brain region or equivalent load. E1 and E3 form a local field potential recording electrode pair for acquiring local field potential signals. E3 can serve as a relatively distant recording reference terminal. In benchtop testing or embodiments requiring a shared stimulation circuit terminal and recording reference terminal, E2 and E3 can be selectively shorted externally to the device; in embodiments where they are arranged independently, E2 and E3 remain separate.

[0034] like Figure 3As shown, the stimulation output module 20 includes a constant current source 20a, an H-bridge switching network 20b, and a current setting unit 81. In one specific embodiment, the current setting unit 81 includes a reference voltage source, a digital-to-analog converter (DAC), and a voltage buffer. The DAC generates a current setting voltage based on the digital code written by the microcontroller 60, and the voltage buffer sends this setting voltage to the constant current source 20a. The constant current source 20a includes an operational amplifier, a matching resistor network, and a sampling resistor, and generates a stimulation current of corresponding amplitude based on the current setting voltage. A pulse width modulation filter circuit capable of generating a stable current setting voltage can also replace the DAC, but without changing the subsequent H-bridge state sequence.

[0035] The H-bridge switching network 20b is powered by a single positive high-voltage power supply. This single positive high-voltage power supply can be generated by a boost circuit 80c in the power module 80, for example, by boosting a low-voltage battery voltage to a positive high voltage suitable for stimulating compliance. The H-bridge switching network 20b includes four controlled switches. Figure 3 In the function numbers shown, S1 is the first high-side switch connecting the constant current source output to E1, S2 is the first low-side switch connecting E1 to the low-side reference node VREF, S4 is the second high-side switch connecting the constant current source output to E2, and S3 is the second low-side switch connecting E2 to the low-side reference node VREF. By selecting different diagonal switches to be turned on, the stimulation current can flow through the stimulation load between E1 and E2 in the first direction or in the second direction opposite to the first direction, thereby generating a biphase stimulation current.

[0036] In the first stimulation phase, S1 and S3 are on, and S2 and S4 are off, with the stimulation current flowing through the stimulation load in the first direction. In the phase-interval state, S1 and S2 of the same first bridge arm are on, while S3 and S4 of the second bridge arm are off. Since E2 is disconnected from both the constant current source output terminal and the low-end reference node, there is no closed stimulation current loop between E1 and E2; the output current of the constant current source 20a returns to the low-end reference node via S1, the midpoint of the first bridge arm, and S2, forming a local current path that does not pass through the stimulation load. In the second stimulation phase, S2 and S4 are on, and S1 and S3 are off, with the stimulation current flowing through the stimulation load in the second direction, opposite to the first direction. Other embodiments can form a mirrored local current path in the other bridge arm, provided that the bridge arm connected to the open electrode is disconnected and the constant current source returns to the reference node via the high-side and low-side switches of the other bridge arm.

[0037] In the phase-interval state comparison test of the prototype, a 60-microsecond first stimulation phase, a 10-microsecond phase interval, and a 60-microsecond second stimulation phase were used. In the original state, disconnecting the stimulation load caused the constant current source to lose its effective current path. It was detected that the output node of the constant current source moved towards the compliance voltage limit within the phase interval, and a recovery peak appeared at the beginning of the next stimulation phase. After changing the phase interval state to S1 and S2 being on and S3 and S4 being off, the stimulation load remained open, and the recovery peak no longer appeared. Under a 47 kΩ load and a stimulation current of approximately 200 μA, the load voltage amplitudes of the two stimulation phases measured after the modification were approximately 9.50 V and 9.59 V, respectively. The difference of 0.09 V between the two was less than one quantization step size under the oscilloscope range. This result is used to illustrate that the local current path controls the simulated operating state of the constant current source during the phase interval, rather than outputting additional stimulation current to the stimulation load within the phase interval.

[0038] For electrode equivalent loads containing a large capacitive component, the node potential at the start of the second stimulation phase is also affected by the charge stored in the previous phase. During implementation, the compliance voltage and switching withstand voltage should be selected based on the stimulation current, phase width, phase interval, and discharge time, and the node transient should be verified to not exceed the allowable range of the stimulation output module 20 and the neural signal acquisition module 30.

[0039] In the stimulation-end state, the H-bridge switching network 20b can provide a discharge path, a bias path, or a combination of both for the stimulation load. The discharge path releases residual charge on the stimulation load, while the bias path returns the common-mode potential of the stimulation load to a predetermined range. In one specific embodiment, the low-end reference node VREF of the H-bridge has a reference potential of approximately 0.9 volts, corresponding to the +0V9 network in the prototype schematic; at stimulation end, at least S2 and S3 are briefly turned on, connecting both E1 and E2 to VREF, thereby forming... Figure 5 The stimulus shown indicates the end of the discharge state, which is also the short-circuit discharge state.

[0040] After the stimulation-end discharge is completed, the H-bridge switch network 20b enters a non-stimulation recording bias state. In this state, the constant current source 20a is turned off or set to zero, and only the low-side switch S3 connected to the second electrode connection terminal E2 is turned on. The upper-side switches S1 and S4 and another low-side switch S2 are turned off, allowing E2 to connect to the low-end reference node VREF of the H-bridge via S3. Since E2 is electrically coupled to the tissue containing E1 and E3 through the electrode tissue interface, this connection provides a DC reference path for the tissue relative to VREF, thereby guiding the common-mode potential at E1 and E3 to the input common-mode range allowed by the neural signal acquisition module 30. Because the other three switches are turned off, there is no complete stimulation current loop formed by the stimulation output module 20 between E1 and E2. After the tissue common-mode potential and the recording front-end input node stabilize, the input blanking switch 31 is closed to resume recording. This non-stimulation recording bias state is different from the stimulation-end discharge state where S2 and S3 are turned on simultaneously; the former is used for tissue common-mode bias during continuous recording, while the latter is used for residual charge release and rapid stabilization after stimulation.

[0041] The load voltage sampling module 40 is used to acquire a scaled voltage related to the voltage across the stimulation load. Figure 3 In the specific embodiment shown, the load voltage sampling module 40 is connected to the detection node VLOAD between the output of the constant current source 20a and the upper end of the H-bridge switching network 20b, instead of directly bridging E1 and E2 for differential sampling. VLOAD is first scaled by a voltage divider network, then buffered by a unity-gain buffer, and finally fed into the analog-to-digital converter 50 of the microcontroller 60 via a low-pass filter network. The prototype uses a 200 kΩ upper voltage divider resistor and a 10 kΩ lower voltage divider resistor, with a nominal voltage division factor of 1 / 21. The buffer output is then low-pass filtered. In the first stimulation phase, one end of the stimulation load is connected to the approximately 0.9V reference node VREF via a low-side switch, and the other end is connected to VLOAD via a high-side switch; in the second stimulation phase, the connection relationship between the two ends is reversed. The load voltage sampling module 40 can also be configured with current limiting and input protection to ensure that the scaled voltage is within the input range of the analog-to-digital converter 50.

[0042] like Figure 4 As shown, the neural signal acquisition module 30 includes an input blanking switch 31, a DC servo preamplifier 32, and a gain and filter circuit 33. The input blanking switch 31 is located between the electrode interface 10 and the DC servo preamplifier 32. During non-stimulation periods, the input blanking switch 31 is turned on, allowing the local field potential signal of the recording electrode pair to enter the DC servo preamplifier 32; during the stimulation window, the microcontroller 60 controls the input blanking switch 31 to be turned off, preventing stimulation artifacts from directly entering the DC servo preamplifier 32.

[0043] The input blanking switch 31 can be used in conjunction with the bias hold network 34. In the prototype, two differential input nodes following the input blanking switch 31 each have a 1 nanofarad hold capacitor connected to ground. While the input blanking switch 31 is open, the hold capacitors limit the rate of change of the input node potential caused by switch-injected charge, leakage current, and parasitic coupling. The hold capacitors themselves do not set the DC bias of the input node; the final bias of the input node is determined by the electrode organization reference path, the preamplifier input characteristics, and the DC servo reference. Other embodiments may employ a high-impedance bias network connected to a predetermined bias potential or a combination of hold capacitors and bias networks.

[0044] The DC servo preamplifier 32 includes an instrumentation amplifier and a servo integrator. In the prototype, the differential gain of the instrumentation amplifier is approximately 201. The servo integrator detects the low-frequency or DC component of the instrumentation amplifier output and applies a correction signal to the reference terminal of the instrumentation amplifier via an attenuation network, bringing the output baseline back to near the reference potential. The integration network uses resistors and capacitors to form a high-pass cutoff frequency of approximately 0.23 Hz to suppress electrode DC bias and slow drift, while preserving the local field potential bandwidth. When using different instrumentation amplifier topologies, the servo correction signal can also be applied to the feedback terminal or an equivalent summing node at the input terminal, as long as the resulting low-frequency negative feedback brings the output DC component back to the reference potential. The gain and filtering circuit 33 provides further gain and analog anti-aliasing filtering after the preamplifier; the prototype uses a fourth-order low-pass filter with a cutoff frequency of approximately 10 kHz, and the 100 Hz bandwidth limitation of the local field potential is achieved by digital filtering after sampling.

[0045] The analog-to-digital converter 50 can be an internal analog-to-digital converter of the microcontroller 60 or an external analog-to-digital converter. In the prototype, the same analog-to-digital converter 50 acquires local field potential signals and load voltage samples in a time-division multiplexing manner. Before entering the stimulus conversion window, the microcontroller 60 stops sending continuous sampling triggers to the local field potential channel and ends the current continuous sampling; then, it selects the channel corresponding to the load voltage sampling module 40 and completes a single sampling at predetermined times during the first and second stimulus phases; after the stimulus conversion window ends, it reselects the local field potential channel and resumes continuous sampling. The continuous sampling counter continues to count during channel switching to determine the number of missing samples.

[0046] like Figure 5As shown, the microcontroller 60 performs stimulation control and signal acquisition according to a hardware timing reference. The microcontroller 60 first writes the stimulation current amplitude and pulse parameters, pauses continuous acquisition of local field potentials, and disconnects the input blanking switch 31 before the stimulation window begins. The hardware timing event then initiates the H-bridge switch sequence. The first load voltage sampling trigger occurs near the end of the first stimulation phase and when the current has stabilized; the second load voltage sampling trigger occurs near the end of the second stimulation phase and when the current has stabilized. After the second stimulation phase ends, the H-bridge switch network 20b first enters the stimulation end discharge state, and then enters the non-stimulation recording bias state where only S3 is on. After a predetermined recovery time, once the residual charge of the stimulation load, the tissue common-mode potential, and the preamplifier input node have stabilized, the input blanking switch 31 closes, the analog-to-digital converter 50 switches back to the local field potential channel, and continuous acquisition resumes.

[0047] The hardware timing reference can be implemented using a timer from the microcontroller 60, a pulse width modulation peripheral, a programmable hardware event interconnect, and a direct memory access controller. In the prototype, the four control words of the H-bridge are stored in memory in 1-microsecond time slices, and the direct memory access controller sequentially provides control words to the pulse width modulation peripheral. Timer comparison events directly trigger the pulse width modulation sequence start, input blanking control, and analog-to-digital converter sampling tasks via the programmable hardware event interconnect. The wireless communication task only updates the parameters used for the next stimulus and does not directly generate stimulus phase edges or load voltage sampling edges.

[0048] The microcontroller 60 can obtain impedance-related parameters based on the load voltage sample and the set stimulation current. Let the scaled voltage measured by the analog-to-digital converter be VS, the zero-point bias of the analog-to-digital conversion link be B, and the calibrated voltage divider coefficient be K. The estimated voltage of the detection node can be expressed as VD = (VS - B) / K. Taking the voltage of E1 relative to E2 as positive, the estimated load voltage of the first stimulation phase can be expressed as VL+ = VD - VREF - VSW, and the estimated load voltage of the second stimulation phase can be expressed as VL- = -(VD - VREF - VSW). The processing of B and VREF constitutes the reference node bias correction, the choice of sign constitutes the stimulation phase sign correction, and VSW represents the voltage drop correction amount of the corresponding on / off switch and connection path. When high absolute accuracy is not required, VSW can be incorporated into the system calibration bias. The impedance-related parameters can be expressed as Zr = |VL| / I, where I is the set stimulation current or the calibrated measured stimulation current. The sampling time is preferably located in the stable range at the end of the phase. Zr reflects the combined voltage-current relationship of the electrode, tissue, connection, and switching path at the sampling time, and may include electrode polarization and transient components. It is not equivalent to the complete complex impedance obtained by frequency sweep. This parameter can be used for trend comparison under the same electrode and the same timing conditions.

[0049] For local field potential data, the microcontroller 60 acquires raw data at a first sampling rate and determines the missing or discarded sample intervals based on the continuous sampling counter and the stimulation window timestamp. The prototype saves the last valid sample before the interruption and the first valid sample after recovery, and performs linear interpolation between the two according to the number of missing samples. For smaller gaps, the last valid sample before the gap can be repeated until the end of the gap to form an endpoint-preserving reconstruction. Optional spline interpolation uses at least two valid samples on each side of the gap to establish a piecewise polynomial. For periodic stimulation, multiple stimulation windows can be aligned according to hardware timing events to obtain an artifact template, which is then used to replace or correct the corresponding intervals; or an autoregressive model can be fitted using valid samples before and after the gap to estimate the missing values. Regardless of the reconstruction method used, the original gap location and the processing method can be saved with the data to avoid mistaking estimated samples for directly measured samples. The processed local field potential data is then digitally filtered and downsampled to form a data stream with a second sampling rate.

[0050] In one specific embodiment, the first sampling rate is 20 kSPS. The linearly interpolated data is passed through a fourth-order Butterworth digital low-pass filter with a cutoff frequency of 100 Hz, and then decimated by a factor of 40 to form a 500 SPS data stream. When the sampling rate or decimation factor is changed, the digital low-pass filter should be adjusted accordingly so that its cutoff frequency is lower than the Nyquist frequency after downsampling.

[0051] like Figure 6 As shown, the power module 80 includes a battery pack 80a, a low-voltage power supply path 80b, and a boost circuit 80c. The battery pack 80a can be a series of button-type zinc-air batteries, such as two PR41 zinc-air batteries, or a lithium battery, silver oxide battery, or other lightweight power source. The low-voltage power supply path 80b can directly supply the battery voltage to a low-voltage load, or it can power the microcontroller 60, wireless communication module 70, neural signal acquisition module 30, and analog-to-digital converter 50 after voltage regulation or bucking. The boost circuit 80c generates a single positive high-voltage rail and provides a stimulus compliance voltage to the stimulus output module 20; in embodiments where the input blanking switch 31 is powered by the single positive high-voltage rail, the single positive high-voltage rail also powers the input blanking switch 31.

[0052] The power supply module 80 may include an input buffer capacitor, a delayed start circuit, a feedback compensation capacitor, and an enable control circuit. In the prototype, both the input blanking switch 31 and the stimulation output module 20 are powered by a single positive high-voltage rail; therefore, the boost circuit 80c remains enabled as long as local field potential recording is performed. After the wireless connection is disconnected, the microcontroller 60 first stops stimulation and acquisition, and then shuts down the boost circuit 80c. In an embodiment where the input blanking switch 31 is powered by an independent low-voltage power supply, the boost circuit 80c can be enabled only when stimulation is required.

[0053] To minimize the impact on the free movement of small animals, device 100 can employ a flexible circuit board, a thin rigid circuit board, or a combination of rigid and flexible circuit board 90. Battery pack 80a can be fixed to circuit board 90 by a battery holder or mechanically connected to circuit board 90. A battery-powered prototype includes a 0.85-gram main circuit board, a 0.75-gram battery holder, and two PR41 batteries with a total weight of 1.00 gram, with a measured total weight of 2.60 grams.

[0054] like Figure 7 As shown, circuit board 90 includes a power supply and boost area, a high-voltage stimulation area, a load voltage detection area, a microcontroller and wireless area, a low-noise analog acquisition area, an electrode connection area, and an antenna and matching area. The high-voltage switching node and stimulation current path are arranged in separate zones from the microvolt-level local field potential input path. A guard ring is set around the input stage of the low-noise analog acquisition area. The decoupling capacitor of the boost circuit is close to the boost switching node, and the antenna area is cleared. Figure 7 This mainly indicates the aforementioned functional zoning and isolation relationships. As prototype parameters without a limited protection scope, the prototype uses a four-layer flexible circuit board, with the second and third layers being intersecting grid power planes. The board thickness is approximately 0.2 mm, and the external dimensions are approximately 20 mm by 39 mm. The specific stacking structure and shape can be adjusted according to the animal species, connectors, batteries, and fixing structure.

[0055] The microcontroller 60 also implements a safe default state. Upon power-on or reset, the stimulation amplitude is set to zero, the constant current source 20a is turned off, and the H-bridge does not execute the stimulation sequence. The H-bridge is only kept in a non-stimulation recording bias state with only S3 conducting when the boost circuit 80c and the H-bridge switching network 20b are powered and local field potential recording needs to continue. When wireless communication is disconnected, the prototype first terminates the stimulation sequence and turns off the constant current source, then stops local field potential acquisition and turns off the boost circuit 80c. After the boost circuit is turned off, the unpowered H-bridge is no longer described as being able to continuously provide tissue bias. If another embodiment requires continued local recording after wireless disconnection, the H-bridge is kept powered and the non-stimulation recording bias state with only S3 conducting is maintained. The constant current source is only allowed to output a non-zero current after receiving stimulation parameters that have passed range verification and a single valid enable command. The microcontroller 60 can also limit the maximum stimulation amplitude, maximum pulse width, maximum duty cycle, and maximum continuous stimulation time.

[0056] The apparatus and method of this invention are not dependent on any specific device model. The microcontroller 60 can be a system-on-a-chip integrating timers, analog-to-digital converters, pulse-width modulation peripherals, and low-power wireless communication functions, or it can work with a standalone wireless communication module. The wireless communication protocol only handles parameter and data transmission and does not alter the stimulation and sampling timing generated by the hardware timing reference. The stimulation load can be a load formed by implanted electrodes and tissue, or it can be a resistor, capacitor, saline bath, or equivalent electrode model used in bench testing.

[0057] The above embodiments illustrate the structure, state sequence, and working process of the present invention. The scope of protection of this application is determined by the claims. Substitutions in device models, numerical parameters, and module implementation methods should be made while still satisfying the circuit connection relationships, switching states, and synchronous control relationships defined in the claims.

Claims

1. A head-mounted neural stimulation recording device for small animals, characterized in that, It includes an electrode interface, a stimulation output module, a nerve signal acquisition module, a load voltage sampling module, an analog-to-digital converter, and a microcontroller, all mounted on the same circuit board suitable for fixing to the head of a small animal. The electrode interface includes a first electrode connection terminal, a second electrode connection terminal, and a third electrode connection terminal. The first electrode connection terminal and the second electrode connection terminal form a stimulation electrode pair for connecting the stimulation load, and the first electrode connection terminal and the third electrode connection terminal form a local field potential recording electrode pair. The stimulation output module includes a constant current source and an H-bridge switching network. The constant current source is powered by a single positive high voltage power supply relative to a reference node. The H-bridge switching network includes a first bridge arm and a second bridge arm. The first bridge arm includes a first high-side switch connected between the output terminal of the constant current source and the first electrode connection terminal, and a first low-side switch connected between the first electrode connection terminal and the reference node. The second bridge arm includes a second high-side switch connected between the output terminal of the constant current source and the second electrode connection terminal, and a second low-side switch connected between the second electrode connection terminal and the reference node. The neural signal acquisition module includes an input blanking switch disposed between the local field potential recording electrode pair and the preamplifier; The load voltage sampling module is connected to the detection node between the constant current source output terminal and the H-bridge switch network; The microcontroller is configured to control a first high-side switch, a first low-side switch, a second high-side switch, and a second low-side switch via a switch sequence triggered by a hardware timer, the switch sequence comprising: In the first stimulation phase, the first high-side switch and the second low-side switch are turned on so that the stimulation current flows through the stimulation load in the first direction; In a phase-interval state, the first high-side switch and the first low-side switch are turned on and the second high-side switch and the second low-side switch are turned off, so that no stimulation current flows through the stimulation load because the second electrode connection terminal is in an open circuit state. At the same time, the output terminal of the constant current source is connected to the reference node through the first high-side switch and the first low-side switch to form a local current path that does not pass through the stimulation load. A second stimulation phase is initiated, turning on the first low-side switch and the second high-side switch, so that the stimulation current flows through the stimulation load in a second direction opposite to the first direction; and In the non-stimulation recording bias state, the second low-side switch is turned on and the other three switches are turned off, so as to bias the tissue common-mode potential to the input common-mode range allowed by the neural signal acquisition module through the second electrode connection terminal; The microcontroller is further configured to synchronously control the input blanking switch based on the hardware timer to open within a stimulation window that covers at least the first stimulation phase, the phase interval state, and the second stimulation phase, close after entering the non-stimulation recording bias state, and trigger the analog-to-digital converter to acquire voltage samples of the detection node at a predetermined time in at least one of the first stimulation phase and the second stimulation phase.

2. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The switching sequence further includes a stimulation end discharge state located between the second stimulation phase and the non-stimulation recording bias state, in which at least the first low-side switch and the second low-side switch are turned on, so that both the first electrode connection and the second electrode connection are connected to the reference node to release the residual charge of the stimulation load.

3. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The constant current source includes an operational amplifier, a matching resistor network, and a sampling resistor. The stimulation output module also includes a digital-to-analog converter and a buffer. The constant current source generates a stimulation current based on a set signal output by the digital-to-analog converter through the buffer.

4. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The neural signal acquisition module also includes a holding capacitor or bias holding network connected to the differential input node after the input blanking switch, for limiting the potential change of the preamplifier input node during the period when the input blanking switch is open.

5. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The preamplifier is a DC servo preamplifier, which includes an instrumentation amplifier and a servo integrator. The servo integrator applies a correction signal to the reference terminal, feedback terminal, or input equivalent node of the instrumentation amplifier based on the low-frequency or DC component output by the instrumentation amplifier.

6. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The load voltage sampling module includes a voltage divider network, a buffer, and a filter network connected sequentially between the detection node and the analog-to-digital converter; the microcontroller is configured to perform voltage divider ratio correction, reference node bias correction, and phase sign correction on the voltage sample according to the stimulation phase corresponding to the sampling time, and obtain impedance-related parameters based on the corrected voltage sample and the stimulation current.

7. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The analog-to-digital converter acquires local field potential signals and voltage samples of the detection node in different time slices; the microcontroller pauses continuous acquisition of local field potential before acquiring the voltage samples, switches the analog-to-digital converter to the load voltage sampling module, and switches back to the local field potential sampling channel and resumes continuous acquisition of local field potential after the voltage samples are acquired.

8. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The microcontroller is configured to mark, remove, linearly interpolate, preserve endpoints, interpolate splines, replace templates, estimate models, or combinations thereof for local field potential samples that are missing or affected by input blanking within the stimulation window, and then perform digital filtering and downsampling on the processed local field potential data.

9. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, The hardware timer is connected to or cooperates with at least one of a pulse width modulation peripheral, a programmable hardware event interconnect, and a direct memory access controller to trigger the switch sequence, the input blanking switch, and the analog-to-digital converter without relying on wireless communication task scheduling.

10. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, It also includes a wireless communication module, which is used to send local field potential data, the voltage sample, impedance-related parameters or device status obtained from the voltage sample and the stimulation current, and to receive stimulation current amplitude, pulse width, stimulation frequency, input blanking window, sampling rate or stimulation start / stop parameters.

11. The small animal head-mounted neural stimulation recording device according to claim 1, characterized in that, It also includes a battery and a boost circuit for generating the single positive high voltage power supply; the electrode interface, stimulation output module, nerve signal acquisition module, load voltage sampling module, analog-to-digital converter and microcontroller are mounted on a flexible circuit board, a rigid-flexible circuit board or a small rigid circuit board, and the total mass of the device in the battery-powered configuration is less than 4 grams.

12. The small animal head-mounted neural stimulation recording device according to claim 10, characterized in that, The microcontroller is configured to shut down the constant current stimulation output upon power-on, reset, or disconnection of wireless communication; to put the H-bridge switching network into the non-stimulation recording bias state while the power supply to the H-bridge switching network remains effective and local field potential acquisition continues; to shut down the single positive high-voltage power supply when local field potential acquisition stops; and to allow the output of stimulation current only upon receiving valid stimulation parameters and an enable command.

13. A synchronous control method for a small animal head-mounted neural stimulation recording device, the device comprising a first electrode connection terminal, a second electrode connection terminal, and a third electrode connection terminal, a constant current source powered by a single positive high-voltage power supply relative to a reference node, an H-bridge switch network composed of a first bridge arm and a second bridge arm, an input blanking switch, a load voltage sampling module, an analog-to-digital converter, and a microcontroller, wherein the first electrode connection terminal and the second electrode connection terminal form a stimulation electrode pair, and the first electrode connection terminal and the third electrode connection terminal form a local field potential recording electrode pair, the input blanking switch being disposed between the local field potential recording electrode pair and a preamplifier, the high-side switch of the first bridge arm being connected between the output terminal of the constant current source and the first electrode connection terminal, and its low-side switch being connected between the first electrode connection terminal and the reference node, the high-side switch of the second bridge arm being connected between the output terminal of the constant current source and the second electrode connection terminal, and its low-side switch being connected between the second electrode connection terminal and the reference node, and the load voltage sampling module being connected to a detection node between the output terminal of the constant current source and the H-bridge switch network, characterized in that... The method includes: The H-bridge switching network is triggered by a hardware timer to enter the first stimulation phase, which turns on the high-side switch of the first bridge arm and the low-side switch of the second bridge arm, so that the stimulation current flows in the first direction through the stimulation load between the first electrode connection terminal and the second electrode connection terminal. After the first stimulation phase, the phase-interval state is entered, the high-side switch and low-side switch of the first bridge arm are turned on and the high-side switch and low-side switch of the second bridge arm are turned off, so that the stimulation load is kept open, and the constant current source is connected to the reference node through the first bridge arm to form a local current path that does not pass through the stimulation load. After the phase-interval state, the second stimulation phase is entered, and the low-side switch of the first bridge arm and the high-side switch of the second bridge arm are turned on, so that the stimulation current flows through the stimulation load in a second direction opposite to the first direction; After the second stimulation phase, a non-stimulation recording bias state is entered, where only the low-side switch connected to the second electrode connection is turned on while the other three switches are turned off, so as to bias the tissue common-mode potential to the input common-mode range allowed by the preamplifier via the second electrode connection. Based on the hardware timer, the input blanking switch is opened within a stimulation window that at least covers the first stimulus phase, the phase interval state, and the second stimulus phase, and closes after entering the non-stimulation recording bias state; and At a predetermined time in at least one of the first stimulus phase and the second stimulus phase, the analog-to-digital converter is triggered to acquire voltage samples of the detection node via the load voltage sampling module.

14. The synchronization control method according to claim 13, characterized in that, Between the second stimulation phase and the non-stimulation recording bias state, the H-bridge switch network is brought into a stimulation-end discharge state so that both the first electrode connection and the second electrode connection are connected to the reference node; and, when acquiring the voltage sample, the continuous acquisition of local field potential is paused, the analog-to-digital converter is switched to the load voltage sampling module, and after the sampling is completed, it is switched back to the local field potential sampling channel.

15. The synchronization control method according to claim 13, characterized in that, The device further includes a wireless communication module; it performs voltage division ratio correction, reference node bias correction, and stimulus phase sign correction on the voltage sample, and obtains impedance-related parameters based on the corrected voltage sample and stimulus current; it marks, removes, or reconstructs local field potential samples that are missing or affected by input blanking within the stimulus window, then performs digital filtering and downsampling, and sends the processed local field potential data and the impedance-related parameters through the wireless communication module.

Images