A Multifunctional, Embedded-based, Bluetooth-enabled, Programmable, Biphasic-waveform Stimulator with Real-time Neural Signal Acquisition

Summary

Here, we describe a new device specifically designed for preclinical deep brain stimulation (DBS) research along with a neural signal recording unit. This device provides extensive flexibility in waveform parameters, including shape, frequency, pulse width, and amplitude, and the recording of neural spikes and local field potentials.

Abstract

The study introduces a lightweight, Bluetooth-enabled deep brain stimulation (DBS) device that is designed specifically for preclinical research in small, freely moving animals. To address the current limitations in waveform flexibility and wireless control, we developed a compact, multifunctional stimulator capable of delivering programmable, charge-balanced biphasic waveforms while simultaneously acquiring neural signals. The system integrates off-the-shelf components for current regulation, signal amplification, and analog-to-digital conversion, all managed by a low-power microcontroller. The assembly includes key steps such as circuit integration on a double-sided 30 x 30 mm PCB, waveform programming via nRF Connect, and validation through saline and load resistance testing. In vitro assessments demonstrated reliable current output across varying load impedances, effective saline operation without waveform degradation, and high-fidelity neural signal recording with a signal-to-noise ratio exceeding 35 dB. These results confirm the device's suitability for closed-loop neuromodulation experiments and lay the groundwork for future translational studies in DBS therapy.

Introduction

This work introduces a novel deep brain stimulator to explore changes at the molecular and behavioral levels in experimental animals. Although various micro-stimulator devices have been discussed in the literature, they vary significantly in size, weight, stimulation capabilities, ability to deliver prolonged stimulation, and the requirement for external power sources. Most rodent devices rely on external power, which limits their utility in studies requiring simultaneous stimulation and behavioral observation¹. While some solutions, such as tethers, magnetic waves, or light sources, attempt to address this issue, each comes with its own set of limitations: tethers restrict movement and behavior, magnetic stimulators are confined to specific environments, and light-based systems are dependent on the lighting cycle².

Most traditional neurostimulation systems for rodents involve tethering the animal to an external stimulator, which not only restricts their freedom of movement but also complicates behavioral testing³. Additionally, the cables can wear out over time, posing a risk of malfunction. While there has been progress in developing portable deep brain stimulation (DBS) systems for rodents, many current devices lack detailed documentation of their circuitry. Often, these devices are too heavy and bulky relative to the rodent's size, and their small batteries are quickly drained by continuous high-frequency stimulation⁴. Some researchers have tried using larger battery packs worn by the animals, but this again limits their natural movements. Furthermore, many devices do not allow for voltage adjustments, which is essential for maintaining consistent stimulation as the impedance changes due to interactions between the electrode and tissue over time⁵. There is a growing need for miniaturized, wireless, and low-power devices that are suitable for eventual human application, posing a new challenge for the development of next-generation closed-loop brain-machine interfaces (BMIs). Non-human primates or rodents are typically used in BMI research because of their similarities to human brains or the availability of a wide range of experimental models⁶. However, the development of lightweight, wireless micro-stimulators for rodents like mice is more challenging due to their smaller size.

As an initial step toward further development, we concentrated on creating a device for rodents with a multifunctional device with stimulations and recording. Researchers usually take one of two approaches: developing custom integrated circuits (ICs) or using commercial-off-the-shelf (COTS) electronics⁷. While ICs offer benefits such as reduced noise, compact design, and energy efficiency, they require longer development times and are expensive, making them more appropriate for long-term implants rather than short-term studies in animals. Conversely, COTS-based systems are more flexible and can be modified easily during early research stages. However, existing COTS systems with recording and stimulating functions are not ideal for long-term experiments with freely moving small animals⁸. They often offer only single-channel capabilities, are wired, and are not suitable for prolonged behavioral studies. Some systems support wireless data transmission but are limited in real-time processing, making them inadequate for closed-loop studies like spike-timing-dependent plasticity.

To overcome these challenges, we integrated both strategies into our device design. We developed a custom IC capable of continuously recording neural activity with high efficiency, minimal noise, and a compact footprint, reducing interference and cross talk. The stimulator was built using COTS components on custom-made circuit boards connected to a microcontroller that handles signal processing, wireless transmission, and programmable stimulation. This approach has resulted in a low-power, low-noise, compact recording system paired with a programmable microstimulator capable of operating in a closed-loop mode. The microcontroller's flexibility in adjusting the loop configuration and transmitting data wirelessly makes it a versatile tool for preclinical DBS research. The results of in-vitro tests are presented here, demonstrating the system's potential to advance the understanding and application of DBS in preclinical studies and beyond.

Protocol

System architecture
In Figure 1, a low-power microcontroller acts as the core of the system, controlling waveform generation, signal processing, and the protocol of communication. The amplitudes of the waveforms are controlled by the programmable current source used. The system uses an integrated H-Bridge circuit for alternating the current direction, and thus creates the biphasic waveforms required to stimulate neural activity. A microcontroller-aided digital resistor, in combination with a shunt resistor, enables programmable waveform modifications through an I2C interface protocol. To ensure that the current passing through electrodes remains constant and consistent with a predefined set of parameters, a feedback loop is provided through a microcontroller. To ensure reliable power delivery and operational stability, a low-dropout voltage regulator is implemented, supplying a 3.3 V power source to the microcontroller along with additional components. The system's communication with external devices, such as a computer, is facilitated by a USB-to-serial converter, which allows for USB connectivity necessary for data acquisition or system configuration. These neural signals obtained from the electrodes are passed through a stage of amplification and filtering to maximize their quality before being processed through the microcontroller. The amplified and filtered signals can either be forwarded for later analysis or stored for further processing, while the system continues to stimulate the neural tissue with the biphasic waveforms generated. A complete block diagram view of the system is shown in Figure 1.

Hardware and schematics
The hardware system consists of several subsystems that work together to produce biphasic waveforms and capture neural signals in real time. It is designed for wireless, remote operation and incorporates low-power components, dependable communication, and robust signal acquisition. Below is a detailed overview of the hardware architecture. The schematic design is shown in Figure 2; also see the Table of Materials.

Power management unit
The device can be powered by a rechargeable lithium-polymer battery or via USB. The battery supplies power to all components: the microcontroller, DAC, amplifiers, and the BLE module. To ensure efficient and stable operation, a voltage regulator is used to provide a consistent 3.3 V power supply to critical components. This regulator's low dropout feature helps extend battery life. Additionally, a battery management system (BMS) is integrated to monitor and manage the battery charging and to prevent overdischarging⁹. The BMS communicates with the microcontroller to deliver real-time battery updates, which are transmitted to the remote interface via the BLE module. The BMS continuously monitors battery levels and triggers a low-power mode when the charge falls below a specified threshold, disabling non-essential components like the DAC and BLE module to conserve energy.

Signal acquisition and conditioning
To process the low-amplitude weak neural signals captured by the electrodes, the system includes amplification and filtering stages. The design of the recording unit is shown in Figure 3. An operational amplifier is used for low-noise signal amplification and effective filtering, ensuring high signal quality.

Low-noise pre-amplification
The preliminary stage uses an operational amplifier as a low-noise pre-amplifier with a gain factor of 100. Neural signals, typically ranging from microvolts to millivolts, are amplified to levels suitable for further processing. The opamp was chosen because of its low input noise density of 4.5 nV/√Hz at 1 kHz and high precision, making it suitable for the capture of low-amplitude neural signals without distortion. Its ability to produce a clean signal with a high signal-to-noise ratio (SNR) is crucial, as even minor noise can obscure important data⁵.

Post-amplification and filtering with low-noise opamp
This amplified signal now becomes even more amplified by a gain of 2.5 to reach an ideal level for ADC. A band pass filter is also utilized; it allows a specified frequency band to pass while canceling the interference and noise. The bandwidth of opamp and its zero offset voltage enable appropriate filtering without introducing phase distortion and delay. In general, the bandpass filter is structured in such a way that allows the transmission of frequencies from 1 Hz to 11kHz, embracing most neural signals, while simultaneously dampening low-frequency drift and high-frequency noise, including interference from mains.

Analog-to-digital conversion (ADC)
After filtering, the conditioned signal is digitized by an ADC operating at a 20 kHz sampling rate. The ADC converts the amplified analog signals into discrete digital values for analysis. The opamp ensures the signal quality is preserved, free from noise and distortion. The high sampling rate of 20 kHz captures rapid changes in neural activity with precision, avoiding aliasing.

Waveform generation and current control
The system generates biphasic waveforms using an H-Bridge circuit, allowing current to flow in both directions through the electrodes and creating alternating positive and negative phases necessary for neural stimulation.

Microcontroller control
The microcontroller controls the H-Bridge based on PWM signals. It controls both the frequency and amplitude of the biphasic signal through adjustment in the PWM duty cycle. The current digital-to-analog converter model behaves as a programmable current source, allowing the waveform amplitude to be controlled precisely. This DAC produces an analog signal that adjusts the current flowing through the H-Bridge, thereby ensuring adherence to the specified current parameters.

Feedback loop
A current-sensing resistor placed in series with the electrodes lets the system monitor how much current it is running using an ADC. The microcontroller can make real-time adjustments to the output of the DAC and its waveform by using feedback from the system combined with a PID controller to maintain the specified current levels.

Wireless communication
The BLE module is used in the system to allow communication of wirelessly transmitted data through BLE communication. This module uses the UART protocol for real-time as well as bidirectional communication between the apparatus and a remote user interface. In this design, the BLE module is connected directly to the microcontroller to achieve low-latency integration within wireless data transmission, including neural signals, current settings, and system status, while receiving user commands for adjustment of waveform parameters like duration, amplitude, and frequency. However, the module has low power consumption, thus it is quite suitable for portable neural devices. The module has a communication range of ~30 m, which is ideal for clinical or lab environments.

Remote control and monitoring interface
The system is wirelessly controlled using an app, nRF Connect, which is available on Android and IOS platforms. The terminal communicates with the BLE module using the UART protocol, enabling users to monitor neural activity and to configure waveform parameters in real time. The features include waveform configuration that can adjust waveform type, current amplitude, frequency and phase duration; and wireless programming unit that monitors system configuration and firmware updates can be transmitted wirelessly to the microcontroller, enabling easy customization without physical access to the hardware.

Software
The control software is essential for regulating the device's functionality and operational parameters. It uses the BLE UART protocol for wireless communication between external control interfaces and the device. This setup allows real-time programming of current-controlled biphasic waveforms and other key parameters like frequency and pulse width. Bluetooth-compatible mobile application connect provides a user-friendly interface for sending commands and adjusting parameters. The system can execute preprogrammed stimulation protocols or accept user-defined commands in real time, all while ensuring low power consumption.

Final design and PCB fabrication
We designed this device using commercial PCB design software. It can be used as a backpack for animals. PCB is double-sided with 0.4 mm thickness and dimensions of 30 x 30 x 1 mm. It only weighs 9 g; the final product is shown in Figure 4.

Operational modes
The device has two operational modes, each mode optimized for specific neural applications: stimulation mode and recording mode. These modes are controlled wirelessly through the BLE UART protocol, enabling communication with a mobile device.

Artifacts rejection: As the device works in two modes -- stimulation mode and recording mode -- it is important to ensure artifact rejection is incorporated so that the recorded data reflects genuine brain activity and not electrical noise from stimulation. During recording, the stimulation module is temporarily turned off for a brief time window, usually milliseconds, to avoid stimulation interference. This prevents the recorder from picking up the high-voltage stimulation artifacts.

Stimulation mode: This mode generates current-controlled biphasic waveforms for neural stimulation. The control software dynamically adjusts parameters like pulse amplitude, pulse width, frequency, and inter-pulse intervals. The H-Bridge topology is able to provide very fine control over the direction of the current and shape of the signal waveform. preventing tissue damage during long-term stimulation by minimizing residual charge. The current source permits very fine adjustment in output current to µA resolution.

Recording mode: This mode involves capturing neural signals via electrodes. The signals, normally in the micro-volt (µV) range, are amplified using low-noise amplifiers and filtered through a band pass filter to remove the undesirable noise. The filtered signals are subsequently digitized by an ADC with a sampling rate of 20 kHz and transmitted wirelessly for analysis. Impedance monitoring ensures continuous assessment of electrode contact quality, which is essential for accurate signal capture. The device supports closed-loop operation, enabling real-time adjustment of stimulation parameters based on recorded neural activity, making it suitable for brain-computer interfaces (BCIs), neuroprosthetics, and other closed-loop neuromodulation applications.

Results

The device underwent different verification and evaluation tests, and the product was tested for its feasibility in five stages: Output waveform evaluation, current output test, saline test, voltage drop test, and recording test.

Output waveform evaluation
The primary test of the device was to check the shape and mode of the waveform using a 1 kΩ resistor connected across the device. It has successfully delivered a biphasic waveform with interphase delay in a digital sto...

Discussion

This multifunctional miniature device marks a meaningful step forward in the design of integrated systems for both neural stimulation and signal acquisition. Unlike existing platforms such as those described by Angotzi et al.⁵ and Melo-Thomas et al.⁸, which emphasize either multichannel recording or stimulation, our system uniquely combines programmable, current-controlled biphasic stimulation with real-time neural signal recording in a compact, wireless form factor. T...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
CP2102 USB	Silicon labs		provides a complete plug and play interface solution that includes royalty-free drivers
DMHC30025LSD H-Bridge circuit	DIODES		This new generation complementary MOSFET H-Bridge features low on-resistance achievable with low gate drive.
microcontroller MDBT42/STM32G491CCU6	STMicroelectronics		ARM Microcontrollers - MCU Mainstream Arm Cortex-M4+ MCU 170 MHz with 256 kbytes of Flash memory
Nordic BLE module MDBT42Q	Raytac Corporation		Allow communication of wirelessly transmitted data through BLE communication
OPA322 operational amplifier	TEXAS INSTRUMENTS		Single, 5.5 V, 20 MHz, zero-cross low-noise (6nV/√Hz) RRIO operational amplifier
Tina Ti Software	TEXAS INSTRUMENTS		Simulates frequency response
TLV1117 15 V, 800 mA, Adjustable and Fixed Linear Voltage Regulator	TEXAS INSTRUMENTS		It is a linear voltage that provides up to 800mA of output current with a supported input voltage range from 2.7 V to 15 V.