Protocol to design and implement scalable wireless network system for high-throughput, automated behavioral and circuit neuroscience in mice

Summary

Here, we present a protocol for implementing scalable, modular hardware and software infrastructure to remotely operate programmable miniaturized networks of wireless neural devices in mice. We describe steps for fabricating remote control module (RCM) hardware, software setup, stereotaxic surgery, and probe implantation. We then detail procedures for locomotor, food intake, and social interaction assays. These techniques help enhance automation and throughput for studies of the neurophysiological underpinnings of behavior.

For complete details on the use and execution of this protocol, please refer to Qazi et al.¹

Graphical abstract

Highlights

• •Instructions for fabrication of wireless μ-ILED probes and remote control modules
• •Guidance on configuring local piconet and global internet control of the wireless system
• •Steps for setting up Raspberry Pi web control, scheduling commands, and logging data
• •Steps for implementing wireless optogenetic experiments in laboratory animals

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Here, we present a protocol for implementing scalable, modular hardware and software infrastructure to remotely operate programmable miniaturized networks of wireless neural devices in mice. We describe steps for fabricating remote control module (RCM) hardware, software setup, stereotaxic surgery, and probe implantation. We then detail procedures for locomotor, food intake, and social interaction assays. These techniques help enhance automation and throughput for studies of the neurophysiological underpinnings of behavior.

Before you begin

The protocol below describes the specific steps to create and use a remotely programmable, globally accessible hardware and software infrastructure we call the “Wireless Network for Behavioural Neuroscience” (WNBN). This infrastructure enables remote, scalable, modular, and chronic high-throughput neuroscience studies. In our recent study, we used the WNBN ecosystem for large-scale control of freely behaving rodents and demonstrated the utility of this approach in established models of behavioral and circuit neuroscience, including locomotor, food intake, and social interaction assays.¹

In this protocol, we describe a step-by-step guide to recreate this network of software and hardware tools for neuroscience experimentation. Our protocol will benefit other neuroscience laboratories seeking to coordinate a “hands off” approach to behavioral experimentation. Such an approach can help reduce human presence that may lead to observer effects impacting in vivo studies.^2,3 The WNBN uses a custom-designed user interface and internet server, Bluetooth mesh technology, and affordable consumer hardware. These components allow remote access to simultaneously and selectively control multiple devices and support scalable, high-throughput studies. This approach is highly flexible and customizable in terms of both hardware and potential applications, but this protocol focuses on the behavioral paradigms and experiments demonstrated in our associated study. While this protocol focuses on the Bluetooth Low Energy (BLE) devices used in our previous study¹ and conventional laboratory hardware, it is immediately compatible with other BLE-enabled devices such as 3D-printed optogenetic probes,^4,5 mechanically transformative probes,⁶ and devices with wireless fluid and electrical stimulation capabilities,^7,8 as well as a wide variety of commercially available hardware and accessories.

Before beginning this protocol, ensure that the following preparations are completed:

 Prepare a computer or mobile device with Bluetooth capability and internet access for local or global remote control.

 Download and install the required firmware, control software, and user interfaces described in this protocol.

 Confirm availability of stereotaxic surgical equipment and behavioral testing setups appropriate for the intended experiments.

As such, the time estimates in this protocol may vary depending on factors such as differences in mouse husbandry conditions, as well as those dictated by the experimental demands of each application.

Innovation

Several recent efforts have introduced neuroscience techniques to increase experimental throughput.^{9,10,11,12,13} Many of these approaches are limited in their ability to scale, adapt, and customize to experimental needs. While some such studies only focus on specific behavioural or physiological domains,^11,12,13 others have largely focused on the software component to integrate Internet of Things-based methods with biological applications.¹⁴ The WNBN described in this protocol overcomes these limitations by facilitating remote, scalable, and customizable high-throughput neuroscience studies. In particular, the protocol enables highly customizable experimental operations for large scale in vivo manipulations in freely behaving animals. The ability to use the WNBN in homecage environments extends these applications to the rapidly developing field of studying laboratory animals within their housing, helping to limit experimenter interactions with the test subjects. Furthermore, this protocol describes one approach for monitoring environmental sensors, but this sensing ability can be customized and extended to other cage-based sensors as well.^{15,16,17,18,19,20,21,22} Altogether the WNBN increases the throughput and flexibility of wireless, homecage-based experiments beyond previous applications.

Institutional permissions

All mouse procedures are approved by the Animal Care and Use Committee of Washington University and conformed to U.S. National Institutes of Health (NIH) guidelines. Investigators should receive approval from the appropriate institutional entities for any animal experimentation described here.

Breeding and husbandry of mice

Timing: 12 weeks

Although these mouse lines are discussed in this protocol, other species and strains can easily be incorporated. For example, we have previously extended this protocol for use in Sprague-Dawley rats.¹

• 1.Obtain AgRP-IRES-Cre²³ (RRID:IMSR_JAX:012899, Jackson Laboratories) and Ai32 mice²⁴ (RRID:IMSR_JAX:024109, Jackson Laboratories) for crossing and Channelrhodopsin-2 expression in AgRP-positive neurons.
• 2.Obtain Thy1-ChR2-YFP²⁵ mice (RRID:IMSR_JAX:007612, Jackson Laboratories) for Channelrhodopsin-2 expression in Thy1-positive neurons.
• 3.Obtain DAT-Cre²⁶ mice (RRID:IMSR_JAX:006660, Jackson Laboratories) for injection of AAV5-Syn-FLEX-Chrimson-tdTomato and AAV1-hSyn1-SIO-stGtACR2-FusionRed in dopamine neurons.

Obtaining the adeno-associated virus

Timing: 2 weeks

The blue and red light-sensitive opsins are delivered to the ventral tegmental area (VTA) via intracranial injections. Adeno-associated viruses (AAVs) are a widely used tool for in vivo expression in optogenetics.²⁷

• 4.Obtain AAV5-Syn-FLEX-rc[ChrimsonR-tdTomato] (#62723-AAV5) and AAV1-hSyn1-SIO-stGtACR2-FusionRed (#105677-AAV1) from Addgene.
• 5.Make 5 μL aliquots of the virus and store at −80 °C.

Note: AAV stocks should be thawed on ice immediately before use.

Design and fabrication of the wireless optogenetic system

Timing: 8–10 h (excluding printed circuit board [PCB] lead time; primarily dependent on the parylene coating process)

One of the remote control modules (RCMs) used in this protocol, the wireless optogenetic system, consists of a wireless control module and a wireless optogenetic probe. The module and probe can be easily attached and detached via a connector. The wireless control module receives commands from the remote control center (RCC) wirelessly and drives the microscale inorganic light-emitting diodes (μ-ILEDs) integrated into the wireless optogenetic probe implanted in the brain, delivering optogenetic stimulation to neural circuits according to the desired stimulation parameters.

• 6.
  Construct wireless optogenetic probe.

  • a.
    Design the flexible PCB for the probe substrate (Figure 1A) and place an order.

    Note: Any PCB design software, such as Altium Designer (Altium) or EAGLE (Autodesk), can be used for this procedure. Additionally, PCB fabrication can be outsourced to any vendor (e.g., PCBWay, China). For this 2-layer probe design, a finished copper thickness of 18 μm (=0.5 oz) and a total flexible PCB thickness of 100 μm are recommended. Gerber files (“WNBN_probe_gerber.zip”, 6.8 kB) for the wireless optogenetic probe PCB are available for download from Zenodo.

  • b.
    Solder μ-ILEDs onto the probe substrate.

    • i.
      Apply solder paste (SMDLTLFP10T5, Chip Quik) on the tips of the probe traces (Figure 1B (i)).
    • ii.
      Mount two blue μ-ILEDs (TR2227, Cree; 470 nm; 270 × 220 × 50 μm³) and two orange μ -ILEDs (TCE10-589, Three Five Materials; 589 nm; 235 × 235 × 170 μm³) onto the designated positions with solder paste applied.
    • iii.
      Solder the μ-ILEDs using a soldering iron (temperature ∼215°C).
    • iv.
      Apply a voltage (∼1.5 V) to the two right electrode pads to test the activation of the blue μ-ILEDs (Figure 1B (ii-a)) and to the two left electrode pads for the orange μ-ILEDs (Figure 1B (ii-b)). Confirm that all four μ-ILEDs turn on successfully.
  • c.
    Coat the probe with Parylene C (7 μm-thick) (Figure 1C).

    • i.
      Prepare the Parylene coater by venting the chamber and setting the system according to the manufacturer’s instructions. Wait until the system reaches stable operating conditions.

      Note: Any Parylene coater (e.g., OBT-PC300, Obang Technology) can be used, provided that deposition parameters are adjusted according to the equipment specifications. If in-house coating equipment is unavailable, outsourcing to a commercial Parylene C coating service provider is an acceptable alternative.

    • ii.
      Place the probes inside the deposition chamber.

      Note: Probes should be secured inside a petri dish using Kapton tape to mask the probe connector electrode pads while leaving the probe shank fully exposed.

    • iii.
      Load the required amount of Parylene C dimer into the vaporizer chamber using an aluminum foil container.

      Note: The coating thickness directly affects the waterproof performance of the probe. A typical coating thickness is 7 μm, with a recommended range of 5–10 μm. The deposited thickness depends on the amount of Parylene C dimer and the specifications of the coating system. For example, when using the OBT-PC300 system (Obang Technology), approximately 2 g of Parylene C dimer per 1 μm target thickness is used (e.g., 14 g for a 7 μm coating). Users should adjust the dimer amount based on their specific coater specifications.

    • iv.
      Start the Parylene C deposition process following the manufacturer’s protocol.

      Note: Deposition parameters and operating steps may vary depending on the specific Parylene coating system used. Users should consult the operating manual provided with their instrument. For example, the operating manual for the OBT-PC300 system (Obang Technology) used in this study is not publicly available online but can be obtained directly from the manufacturer upon request.

    • v.
      Remove the probes from the chamber after completion of the deposition process.

      Note: To assess coating thickness, it is recommended to include a flat dummy substrate during deposition. After coating, the Parylene layer on the dummy substrate can be peeled and measured using a profilometer or ellipsometer. A vernier caliper may be used for a rough thickness check if higher-resolution tools are unavailable.

    • vi.
      Perform post-deposition handling according to the specific coater’s operating procedure.

      Note: Post-deposition procedures (e.g., chamber purge, cooling, cold trap cleaning, or re-establishing vacuum) may vary depending on the Parylene coating system. Users should follow the manufacturer’s guidelines for their specific equipment.

  • d.
    Attach a 3-pin female connector to the connector electrode pads and assemble the probe with a 3D-printed probe holder.

    • i.
      Apply solder paste to the connector electrode pads (Figure 1D (i)).
    • ii.
      Mount a 3-pin female connector (M50-3130345, Harwin) onto the electrode pads and solder the connector electrodes using a soldering iron (temperature ∼215 °C) (Figure 1D (ii)).
    • iii.
      Place the probe in a 3D-printed probe holder, apply epoxy (5 Min Epoxy, Permatex), and wait 5–10 min for the epoxy to cure and secure the probe (Figure 1D (iii)).

      Note: The 3D CAD file (“WNBN_probe_holder.STL”, 10.1 kB) for the probe holder is available for download from Zenodo. The probe holder can be printed using any 3D printer, as the choice of resin or filament does not substantially affect the outcome. In this protocol, the B9 Core Med 550 (B9Creations) 3D printer was used with red resin (BioRes – Medical/Wearable, B9Creations). In our associated manuscript,¹ yellow resin (B9R-4-Yellow, B9Creations) was used. However, red resin was chosen in this protocol due to its biocompatibility and compliance with prolonged skin contact. When securing the probe inside the holder with epoxy, care must be taken to prevent epoxy from entering the 3-pin female connector.

• 7.
  Construct wireless control module.

  • a.
    Design the flexible PCB for the wireless module substrate (Figure 2A) and place an order.

    Note: Gerber files (“WNBN_module_gerber.zip”, 10.9 kB) for the wireless control module PCB are available for download from Zenodo. Figure 2B provides an overview of the components used in the wireless control module. The footprints in the PCB layout may require modifications depending on the dimensions of these components. For example, the BLE System-on-chip (SoC) used in the associated previous work¹ (RFD77101, Simblee; 10 × 7 × 2.2 mm³) was discontinued. As a replacement, a new product (EYSHSNZWZ, TAIYO YUDEN; 8.55 × 3.25 × 0.85 mm³) was used, necessitating adjustments to the PCB layout accordingly.

  • b.
    Solder surface-mount device (SMD) components onto the wireless module PCB electrode pads.

    • i.
      Apply solder paste to the electrode pads on either the bottom (Figure 3A (i)) or top (Figure 3B (i)) layer of the PCB.
    • ii.
      Mount the BLE SoC onto the bottom layer electrode pads (Figure 3A (ii)) or mount the SMD components labeled (a) to (f) and (i) in Figure 2B onto the top layer electrode pads (Figure 3B (ii)).
    • iii.
      Place the component-mounted PCBs into a reflow oven (e.g., AS-5060, SMTmax) and solder them with a peak temperature of 170 °C for 60 s (Figure 3C).

      Note: For the recommended temperature profile according to the solder paste type and for instructions on adjusting the temperature settings, please refer to the AS-5060 operating manual provided by SMTmax.

    • iv.
      Repeat Step 7b (i)-(iii) for the opposite side of the wireless module PCB.

      Note: Either the top or bottom layer can be soldered first. However, since the BLE SoC is the only component located on the bottom layer, it is recommended to solder the BLE SoC first before soldering the remaining components on the top layer.

  • c.
    Solder 3- and 2-pin male connectors to the connector electrode pads.

    • i.
      Apply solder paste to the connector electrode pads.
    • ii.
      Mount 3-pin (M50-3630342, Harwin) and 2-pin (M50-3630242, Harwin) male connectors onto the electrode pads and solder the connector electrodes using a soldering iron (temperature ∼215 °C) (Figure 3D).
  • d.
    Enclose the PCB with a battery.

    • i.
      Solder a 2-pin female connector (M50-3130245, Harwin) to the power cable of a lithium polymer (LiPo) battery (GMB-400909, Guangzhou Markyn Battery Co., Ltd.; 25 mAh; 9 × 9 × 4 mm³).
    • ii.
      Position the PCB, with components soldered on both the top and bottom layers, onto the LiPo battery.
    • iii.
      Wrap Kapton tape around the module once horizontally and once vertically, ensuring that the 3- and 2-pin male connectors remain uncovered (Figure 3E).
• 8.
  Upload firmware to the BLE SoC.

  • a.
    Power the wireless module by connecting the 2-pin connectors of the PCB and battery.
  • b.
    Switch the BLE SoC to Bootloader mode.

    • i.
      Launch “nRF Connect for Mobile” (Nordic Semiconductor ASA) application (app), then connect to the chosen RCM to upload the firmware (Figure 4 (i)).
    • ii.
      Send the command character ‘z’ (Figure 4 (ii) – (iv)).
    • iii.
      Confirm that the connection to the RCM has been discontinued (Figure 4 (v)).

      Note: “nRF Connect for Mobile” app is available for free on the Google Play Store and the Apple App Store. Skip Step 8b if this is the first time performing an over-the-air firmware upload on this BLE SoC, so the Bluetooth device name already appears as “BOOTLOADER.”

  • c.
    Perform an over-the-air firmware update for the BLE SoC.

    • i.
      Launch “nRF Connect for Mobile” app and connect to the device in Bootloader mode, which is displayed as “BOOTLOADER” (Figure 5 (i)).

      Note: If the device name does not appear as “BOOTLOADER,” it means the BLE SoC has not entered Bootloader mode. Repeat Step 8b.

    • ii.
      Select the distribution packet file to upload from a storage folder on your smartphone (Figure 5 (ii)-(iii)).

      Note: Distribution packet files (“WNBN_Device01.zip,” . . . ,“WNBN_Device02.zip,”, “WNBN_Device10.zip”, 95.7 kB each) for firmware update are available for download from Zenodo. The firmware contents embedded in each file are identical, with only the registered device name differing. For example, uploading the “WNBN_Device01.zip” file will set the device name of the wireless module to “Device01.”

    • iii.
      Wait until the upload is complete (Figure 5 (iv)). Verify that the device name has been correctly changed to the name corresponding to the newly uploaded firmware (e.g., “Device01,” “Device02,” or “Device10”).

Figure 1.

Construction process of the wireless optogenetic probe

(A) PCB layout of the optogenetic probe. Orange, blue, and red trace/pads indicate the nets for the anode of the blue μ-ILEDs, the anode of the orange μ-ILEDs, and the common ground, respectively.

(B) Apply solder paste to the designated positions for the μ-ILED electrodes (i), then solder four μ-ILEDs onto the traces. Activate the blue (ii-a) and orange (ii-b) μ-ILEDs individually by applying currents to the corresponding electrode pads. Scale bar, 1 mm.

(C) Coat the probes with Parylene C. Inset shows the probes placed in a Petri dish inside the deposition chamber.

(D) Apply solder paste to the connector electrode pads (i), solder a 3-pin female connector (ii), then assemble the probe with a 3D-printed probe holder (iii). Scale bar, 5 mm.

Figure 2.

PCB design and component overview of the wireless control module

(A) PCB layout of the top (left) and bottom (right) layers. Components enclosed by red (a–f), green (g–h), and blue (i–j) boundaries represent passive components, connectors, and active components, respectively.

(B) Component list used in the wireless control module.

Figure 3.

Construction process of the wireless control module

(A) Apply solder paste to the bottom layer pads (i), then mount the BLE SoC (ii). Scale bar, 2 mm.

(B) Apply solder paste to the top layer pads (i), then mount SMD components (ii). Scale bar, 2 mm.

(C) Perform reflow soldering in a reflow oven.

(D) Solder 3- and 2-pin male connectors to the connector pads. Scale bar, 5 mm.

(E) Place the assembled PCB onto the battery (i), then wrap it once horizontally and once vertically using Kapton tape (ii). Inset in the right image shows a cross-sectional view of the completed module. Scale bar, 10 mm.

Figure 4.

Transition process to Bootloader mode

Switch the BLE SoC to Bootloader mode by sending the command character ‘z’. Follow the indicated order (#1–#6).

Figure 5.

Over-the-air firmware update process

Update firmware wirelessly using “nRF Connect for Mobile” (Nordic Semiconductor ASA). Follow the indicated order (#1–#4).

Setup of traditional in vivo behavioral equipment and/or sensor modules

Timing: ∼1 h

The RCMs controllable via WNBN are not limited to custom-made wireless neural devices. WNBN can also be effectively deployed to control conventional equipment commonly used in behavioral experiments. For example, diode-pumped solid-state (DPSS) lasers and conventional LEDs, widely used for optogenetic stimulation, can be controlled via WNBN by integrating driver hardware with a wireless control module. In addition to devices for neural circuit stimulation, monitoring devices can also function as RCMs within the WNBN system. As an example, this protocol uses an atmospheric sensor module for environmental monitoring during behavioral experiments, demonstrating its operation through WNBN.

• 9.
  Setup for DPSS laser-based conventional device integration.

  • a.
    Attach the wireless control module onto the battery pack.

    • i.
      Construct wireless control module by completing Step 7b without integrating the 3-pin and 2-pin male connectors and battery (Figure 6A, right).
    • ii.
      Insert three rechargeable AAA batteries (EBL AAA Rechargeable Ni-MH Batteries 1100mAh, EBL; 1100 mAh) into the battery pack (ZR-BC-9B26AS13A, ZRM&E; 6.2 × 3.5 × 1.5 cm³).
    • iii.
      Attach the module onto the battery pack using double-sided tape (Figure 6A, left).

      Note: While the same LiPo battery (25 mAh) used in Step 7d (i) can be used, a higher-capacity battery is recommended for powering benchtop equipment.

  • b.
    Connect the battery pack and the Bayonet Neill-Concelman (BNC) cable to the module.

    • i.
      Solder the wire leads from the battery pack to the battery electrode pads on the wireless module. Based on the right image in Figure 6B, connect the red lead to the left pad (anode) and the black lead to the right pad (cathode, GND).
    • ii.
      Remove one connector from each of two male-male (M-M) jumper wires, strip the insulation, and solder them to the left pad (anode) and the middle pad (cathode, GND) among the three LED electrode pads, as shown in the right image of Figure 6B.

      Note: Alternatively, the anode jumper wire can be connected to the right pad instead of the left pad. In this case, LED 2 must be driven instead of LED 1.

    • iii.
      Connect the jumper wires attached to the LED electrode pads to the alligator clips of a BNC to alligator clip cable, matching the red (anode) and black (cathode) clips accordingly (Figure 6B, left).

      Note: Ensure that the alligator clips do not touch one another to avoid electrical short circuits.

  • c.
    Integrate the wireless control module with the laser driver.

    • i.
      Connect the BNC cable attached to the wireless module to the Trigger port of the laser driver (BL473T8-100FC, Shanghai Laser & Optics Century Co., Ltd.; 473 nm) (Figure 6C (i)).
    • ii.
      Turn on the power of the battery pack and the laser driver.

      Note: To verify the operation of this setup, activate LED 1 using the local piconet control (Steps 11–12) or the global internet control (Steps 13–17), and confirm that light is emitted from the optical fiber connected to the laser driver (Figure 6C (ii)).

• 10.
  Setup for atmospheric sensor module-based remote data collection.

  • a.
    Connect the atmospheric sensor to the Bluetooth control board.

    • i.
      Solder a 6-pin female connector (DS1023-1∗6S21, Connfly) onto the atmospheric sensor module (SparkFun Atmospheric Sensor Breakout - BME280, SparkFun) (Figure 7A).
    • ii.
      Connect the microcontroller board (Arduino UNO Rev3, Arduino), BLE module (HM-10 BLE Bluetooth V4.0 Module, OEM), and atmospheric sensor module using jumper wires, according to the pin wiring diagram shown in Figure 7B.

      Note: In our manuscript,¹ a different BLE breakout module (RFD77203, Simblee) was used. Due to its discontinuation, a combination of a commonly available microcontroller board and BLE module is demonstrated here instead.

  • b.
    Upload the firmware to the microcontroller board.

    • i.
      Install “Arduino IDE” (Arduino) on the PC and launch the program.
    • ii.
      In the “Tools” tab, select “Arduino UNO” as the board, and upload the “WNBN_Arduino_HM10.ino” code to the Arduino UNO board.

      Note: The Arduino code (“WNBN_Arduino_HM10.ino”, 432 Bytes) for HM-10 BLE module device name registration is available for download from Zenodo.

    • iii.
      Open the Serial Monitor and sequentially send the following commands to set the device name of the HM-10 BLE module:

      >AT

      >AT+NAME<Device’s name>

      >AT+RESET

      Note: For example, to set the HM-10 BLE module’s device name to “Sensor01,” send the command “AT+NAMESensor01.” Upon successful execution, the Serial Monitor will display ‘OK’ in response to the “AT” command, “OK+SetSensor01” in response to the “AT+NAMESensor01” command, and “OK+RESET” in response to the “AT+RESET” command (Figure 7C).

    • iv.
      Install the following three additional libraries via the Arduino IDE Library Manager: Adafruit Unified Sensor (by Adafruit), Adafruit BME 280 Library (by Adafruit), and DHT sensor library (by Adafruit).

      Note: It is recommended to maintain the Adafruit BME 280 Library and DHT sensor library at version 1.0.0, rather than updating to the latest versions.

    • v.
      In the “Tools” tab, select “Arduino UNO” as the board, and upload the “WNBN_Arduino_BME280.ino” code to the Arduino UNO board.

      Note: The Arduino code (“WNBN_Arduino_BME280.ino”, 1.6 kB) for BME280 atmospheric sensor module control is available for download from Zenodo.

  • c.
    Install the atmospheric sensing system inside the mouse cage, ensuring that the sensor module is placed within the cage (Figure 7D).

    Note: The atmospheric sensor must be positioned inside the cage to accurately measure ambient temperature and humidity. To minimize damage caused by animal interference, it is recommended to secure the sensor within 5 cm of the cage ceiling, beyond the typical reach of the mouse. If accessible, mice may bite or physically disturb the sensor module, which can damage the circuitry or cause electrical short circuits, particularly in humid conditions. While a protective enclosure may reduce mechanical damage, enclosing the sensor may compromise the accuracy of temperature and humidity measurements and is therefore not recommended unless necessary.

Figure 6.

Setup process for conventional device integration

(A) Attach a wireless control module (right; Scale bar, 2 mm) – excluding the 3- and 2-pin male connectors and battery – onto a battery pack (left).

(B) Solder a pair of power cables from the battery pack and a pair of jumper wires to the module’s electrode pads (right; Scale bar, 2 mm). Then, connect the alligator clips of the BNC cable to the jumper wires (left).

(C) Connect the BNC cable to the laser equipment (i), then operate the module to emit light through the optical fiber (ii).

Figure 7.

Setup process for the atmospheric sensor module for remote data collection

(A) Solder a 6-pin female connector onto the BME280 atmospheric sensor module. Scale bar, 10 mm.

(B) Pin wiring diagram for connecting the Arduino UNO microcontroller board (left), HM-10 BLE module (middle), and BME280 atmospheric sensor module (right).

(C) Screenshot of the Arduino IDE Serial Monitor during the device name setting process for the HM-10 BLE module.

(D) Install the atmospheric sensing system inside a mouse cage.

Setup of RCC for local piconet control

Timing: <10 min

In the WNBN system, when controlling RCMs via the local piconet method, a mobile device such as a smartphone serves as the RCC. This local piconet control method requires no special setup and offers the advantage of extremely low latency (typically ∼20 ms, with a maximum of ∼50 ms).¹ However, the piconet-based approach also has limitations: the user must be physically present within reliable Bluetooth communication range of the RCMs (typically within ∼20 m in practical indoor environments, depending on obstacles and signal conditions), and the number of RCMs that can be simultaneously and selectively controlled is limited to a maximum of 15 devices (subject to variation depending on the type of mobile device and operating system (OS; e.g., Android or iOS)). The following section describes how to install and use the smartphone app for local piconet control.

• 11.
  Install the smartphone app.

  • a.
    Install the “WNBN LocalPiconet” app on the smartphone using the Android Application Package (APK) file.

    • i.
      Download the APK file required to install the “WNBN LocalPiconet” app and save it to the “Download” folder on the smartphone (Figure 8 (i)).

      Note: The APK file (“WNBN_LocalPiconet_app.apk”, 2.1 MB) is available for download from Zenodo. As it was developed using Android Studio, it is only compatible with devices running the Android OS (e.g., Galaxy smartphones). nRF Connect can also be used for direct one-to-one device control in a sequential fashion from any smartphone.

    • ii.
      From the smartphone’s “Download” folder, run the APK file to install the app (Figure 8 (ii) – (iv)).

      Note: During installation, a warning such as “Unsafe app blocked” may appear. However, the app is safe for the device, and installation can proceed by selecting “Install anyway” (Figure 8 (iii)).

  • b.
    In the device settings, configure the app’s location access to “Allow all the time” (Figure 8 (v)).

    Note: For example, on Galaxy devices, this setting can be adjusted via the following path: Settings → Location → App permissions. If location access is not set to “Allow all the time,” the app may not be able to discover nearby devices during execution.

• 12.
  Conduct an initial test of RCM control using the local piconet.

  • a.
    Launch the “WNBN LocalPiconet” app (Figure 9 (i)).
  • b.
    Pair the smartphone with the selected devices.

    • i.
      On the device list page, tap the magnifying glass icon at the top right corner to scan for nearby Bluetooth devices (Figure 9 (ii)).

      Note: After tapping the magnifying glass icon, it temporarily changes to a black square icon for several seconds (<15 s), indicating that the app is scanning for nearby devices. To stop the scan, tap the black square icon again to return it to the magnifying glass icon.

    • ii.
      From the list of discovered devices, long-press (>1 s) the RCMs to be controlled, then tap the “CONNECT” button to initiate pairing (Figure 9 (iii)).
    • iii.
      Once pairing is successful, the app automatically navigates to the device control page (Figure 9 (iv)). Verify successful pairing by checking that “Connected” appears below the device name.

      Note: If the app closes unexpectedly after tapping the “CONNECT” button in Step 12b (ii), this may indicate that the selected device has a different set of universally unique identifiers (UUIDs) than expected, or that an unexpected error has occurred. The app is designed exclusively for controlling wireless modules used in optogenetic stimulation and can only pair with Bluetooth devices configured with the specific UUIDs (write and read characteristic UUIDs: f3641401-00b0-4240-ba50-05ca45bf8abc). If pairing issues occur, see troubleshooting 1 for guidance.

  • c.
    Send commands to the connected devices and verify their operation.

    • i.
      Select the command button to be sent, then tap the “SEND” button at the top right corner to control the RCMs (Figure 9 (v)).

      Note: Selected buttons will turn gray and can be deselected by tapping them again. The status of each RCM can be monitored via the label displayed below its name (e.g., LED1 / 20Hz).

    • ii.
      To stop LED activation, send the “LED Off” command (Figure 9 (vi)).
    • iii.
      To disconnect from the RCMs, tap the “DISCONNECT” button at the top and confirm that the label below the device name changes to “Disconnected” (Figure 9 (vii)).

Figure 8.

Smartphone app installation process

Install the smartphone app using the apk file for local piconet control. Follow the indicated order (#1–#5).

Figure 9.

Initial test process of smartphone app

Control the RCMs based on local piconet using the smartphone app. Follow the indicated order (#1–#9).

Setup of RCC for global internet control

Timing: ∼3 h (primarily spent during the flashing process of the image [IMG] file)

Global internet control enables authorized users to access the web-based user interface and remotely control RCMs located inside the laboratory from any location with an internet connection. Unlike the local piconet control method, which relies exclusively on Bluetooth communication, the global internet control method transmits commands over the internet while maintaining Bluetooth-based communication between the mini-computer and RCMs within the laboratory. Due to this communication structure, latency is generally higher compared to the local piconet control method (minimum ∼55 ms), and additional setup procedures are required. Nonetheless, global internet control offers significant advantages, including the ability to conduct behavioral experiments independent of time and location, and the scalability to simultaneously and selectively control a larger number of RCMs (demonstrated with 22 devices in the associated research paper). The following section describes the setup and operation of the global internet control system.

• 13.
  Install the mini-computer.

  • a.
    Install “SD Card Formatter” (SD Association) on the PC and format the microSD card using the software (Figure 10).

    Note: Because the IMG file to be flashed in the next step is approximately 60 GB, a microSD card with a minimum capacity of 64 GB is required.

  • b.
    Install “Win32 Disk Imager” (SourceForge) on the PC and flash the IMG file, which contains the OS and pre-installed software, for remote internet control – onto the microSD card using the software (Figure 11).

    Note: IMG files (“WNBN_IMG_base.zip”, 2.0 GB, and “WNBN_SENSOR_IMG_base.zip”, 2.1 GB) are available for download from Zenodo. After downloading, extract the ZIP files before flashing. “WNBN_IMG_base.img” is used for controlling the wireless control module for optogenetic stimulation, whereas “WNBN_SENSOR_IMG_base.img” is used for atmospheric sensor-based data collection. Ensure that the appropriate IMG file is selected according to the intended application. The flashing process typically requires 1–2 h to complete.

  • c.
    Set up the mini-computer (Raspberry Pi 4 Model B, Raspberry Pi) for OS configuration, following the schematic shown in Figure 12.

    • i.
      Insert the microSD card into the mini-computer.
    • ii.
      Insert BLE dongles (BLED112, Silicon Labs) into the USB ports of the mini-computer.
    • iii.
      Establish internet connectivity by connecting a local area network (LAN) cable.
    • iv.
      Connect the mini-computer to a monitor via an HDMI cable.
    • v.
      Enable keyboard and mouse usage by connecting either wired peripherals via USB cables or wireless peripherals via a Bluetooth dongle inserted into a USB port.
    • vi.
      Connect the USB-C power cable to supply power to the mini-computer.

      Note: The order of Step 13c (i) – (v) is flexible; however, the power cable must always be connected last. When shutting down the system, the power cable should be disconnected first to prevent OS corruption. A monitor, keyboard, and mouse are only required during the initial OS configuration; once setup is complete, Step 13c (iv)–(v) can be omitted for global internet control.

• 14.
  Adjust OS settings.

  • a.
    Establish a static IP address.

    • i.
      Open the Terminal on the mini-computer and enter the following command to open the “dhcpcd.conf” file:

      >sudo nano /etc/dhcpcd.conf
    • ii.
      Scroll to the bottom of the “dhcpcd.conf” file and append the following static IP address configuration (Figure 13A):

      >interface eth0

      >static ip_address=<IP Address>/24

      >static routers=<Gateway Address>

      >static domain_name_servers=<Primary DNS Server>

      Note: For example, if the intended IP address is 143.248.175.161, input the following:

      >interface eth0

      >static ip_address=143.248.175.161/24

      >static routers=143.248.175.1

      >static domain_name_servers=143.248.1.177
      In the format ‘<IP Address>/24,’ the ‘/24’ specifies the subnet mask corresponding to 255.255.255.0, which defines how the network and host portions of the IP address are allocated.
    • iii.
      Save the “dhcpcd.conf” file by pressing ‘Ctrl+O,’ then exit the editor by pressing ‘Ctrl+X’ to return to the Terminal.
    • iv.
      In the Terminal, reboot the mini-computer by entering the following command to apply the updated static IP address settings:

      >sudo reboot
    • v.
      After the reboot and the reappearance of the monitor display, verify the network connection by entering the following command in the Terminal and confirming continuous data reception:

      >ping <IP Address>

      Note: Depending on institutional network security policies, device authentication based on the combination of the static IP address and the mini-computer’s MAC address may be required. If necessary, consult institutional officials to complete the authentication process.

  • b.
    Set the UUIDs and BLE address type to match the target RCM.

    • i.
      Open the Terminal on the mini-computer and navigate to the “controller.py” Python source code file by sequentially entering the following commands:

      >pwd

      >cd /srv/www/BoulderBlueServer

      >ls

      >cd BoulderBlueServerWebApp

      >cd bt_controller

      >sudo nano controller.py
    • ii.
      In the “controller.py” file, update both the write and read characteristic UUIDs to match the target RCM’s UUID (Figure 13B).
    • iii.
      Configure the BLE address type within “controller.py” to either public or random, depending on the hardware specifications of the RCM.
    • iv.
      Save the “controller.py” file by pressing ‘Ctrl+O,’ then exit the editor by pressing ‘Ctrl+X.’
    • v.
      In the Terminal, reboot the mini-computer by entering the following command’ to apply the updated UUID and BLE address type settings:

      >sudo reboot

      Note: If using the provided firmware (i.e., “WNBN_Device01.zip,” …, “WNBN_Device10.zip,” “WNBN_Arduino_HM10.ino,” and “WNBN_Arduino_BME280.ino”), Step 14b can be omitted, as the appropriate UUIDs and BLE address type are pre-configured within the “WNBN_IMG_base.img” or “WNBN_SENSOR_IMG_base.img” files. For instance, when controlling a wireless control module for optogenetic stimulation (i.e., using “WNBN_IMG_base.img”), the write and read characteristic UUIDs are pre-set to ‘f3641401-00b0-4240-ba50-05ca45bf8abc,’ and the BLE address type is configured as ‘pygatt.BLEAddressType.random,’ due to the dynamic MAC address behavior of the embedded BLE SoC. Conversely, when collecting atmospheric sensor-based data (i.e., using “WNBN_SENSOR_IMG_base.img”), the UUIDs are set to ‘0000ffe1-0000-1000-8000-00805f9b34f,’ corresponding to the default UUID of the HM-10 BLE module, and the BLE address type is set to ‘pygatt.BLEAddressType.public,’ as the HM-10 utilizes a fixed MAC address.

  • c.
    Back up the updated IMG file reflecting the desired setting using “Win32 Disk Imager” (SourceForge) to simplify future reconfiguration (Figure 14).
• 15.
  Customize the web user interface.

  Note: After powering on the mini-computer, wait at least 5 min before accessing the website to allow all services to initialize. It is also recommended to use the Chrome browser for optimal compatibility.

  • a.
    Log in to the web interface using the default credentials: ID ‘root’ and password ‘chupist2021’ (Figure 15A).

    Note: The website address follows the format “<IP Address>/mouse_viewer.” For example, if the IP address configured in Step 14a (ii) is 143.248.175.161, the website address would be “143.248.175.161/mouse_viewer.” If access to the website fails, see troubleshooting 2 for guidance.

  • b.
    To create a new account, navigate to the following path on the “Admin” page (Figure 15B): Authentication and Authorization → Users → Add.
  • c.
    To configure the list of commands to be sent to the RCM, go to the following path on the “Admin” page and modify the command list according to the specific application (Figure 15C): Blue_Views → Chip commands → Add.
• 16.
  Conduct an initial test of real-time RCM control via a global internet network.

  • a.
    Pair with selected devices on the “Home” page.

    • i.
      After logging into the website, navigate to the “Home” page and click the refresh icon next to the “Discovered Chips” label to scan for nearby Bluetooth devices around the mini-computer (Figure 16 (i)).
    • ii.
      Click the “Actions” button next to the desired RCM name to register the device. Once registered, the devices are automatically paired.
    • iii.
      Verify successful pairing by checking whether the device name appears in green under the “Registered Chips” list (Figure 16 (ii)).

      Note: In the “Registered Chips” list, a green device name indicates successful pairing, while a black name indicates that the device is not currently paired. If pairing issues occur, see troubleshooting 1 for guidance. To deregister or disconnect a device, click the “Actions” button next to its name and select “Deregister” or “Disconnect.” Similarly, to pair a registered but currently disconnected device, click “Actions” and then select “Connect.”

  • b.
    Send commands to connected devices on the “Send” page.

    • i.
      Navigate to the “Send” page and select the devices to receive commands by checking the box to the left of each device name (Figure 17A (i)).
    • ii.
      Click on the device name to display the list of available commands registered in Step 15c. Select one or more commands to send to that specific device.

      Note: Multiple commands can be selected simultaneously. This method is particularly useful when operating multiple μ-ILEDs at once (e.g., simultaneously activating both blue and orange μ-ILEDs). If the same command is to be sent to all selected RCMs, select the command for just one device, then click the “Apply to all selected Devices” button on the right.

    • iii.
      Click the “Send Selected Commands” button at the bottom of the page to transmit the selected commands to the selected RCMs (Figure 17A (ii)). A brief log summarizing the result of the command transmission will appear at the bottom of the page for 2-3s.

      Note: If the command transmission log does not appear or indicates a failure, see troubleshooting 3 for guidance.

  • c.
    Confirm device operation on the “Log” page.

    • i.
      Navigate to the “Log” page and review the command transmission history (Figure 17B (i)).

      Note: To clear the log history, click the “Clear Logs” button on the right side of the page.

    • ii.
      If needed, download the log as an Excel-format file by clicking “Load CSV” in the bottom right corner (Figure 17B (ii)).
• 17.
  Conduct an initial test of scheduled RCM control via a global internet network.

  • a.
    Create scheduled commands on the “Scheduled Commands” page according to the intended use case.

    • i.
      On the “Scheduled Commands” page, click the “Create New Job” button at the top left corner.
    • ii.
      In the “Job ID,” “Chip,” and “Command” fields, enter the name of the scheduled command, select the target device(s), and enter the command to be sent.
    • iii.
      Choose one of the available options – “Schedule Once,” “Schedule Periodic,” or “Schedule Periodic (Advanced)” – and fill out the relevant scheduling parameters accordingly (Figure 18 (i)).

      Note: “Schedule Once” schedules a single execution at a specified future date and time. “Schedule Periodic” allows for repeated execution of the same task starting from a specified date and time until an end date and time. “Schedule Periodic (Advanced)” enables recurring execution at specific intervals defined by year, month, date, hour, minute, and second. For example, in “Schedule Periodic (Advanced),” if “Hour” is set to 12 and “Minute” to 30, the selected command will be executed at 12:30 every day within the specified data range.

    • iv.
      Click the “Submit” button at the bottom to create the scheduled command.
    • v.
      Verify that the scheduled command was created successfully by checking the newly created entry in the “Scheduled Commands” tab (Figure 18 (ii)).
  • b.
    Confirm device operation on the “Log” page.

    Note: If the “Log” page does not display logs for the device operation or indicates a failure, see troubleshooting 3 for guidance.

Figure 10.

MicroSD card format process Format the microSD card using “SD Card Formatter” (SD Association) before flashing the IMG file

Figure 11.

IMG file flash process Flash the OS image and required software to the microSD card using “Win32 Disk Imager” (SourceForge)

Figure 12.

Mini-computer setup schematic Connect the Raspberry Pi (left image) to the internet and a monitor, then configure the OS to match the experimental environment

Figure 13.

OS settings configuration process

(A) Configure a static IP address by editing the “dhcpcd.conf” file.

(B) Set the UUIDs of the target RCM by editing the “controller.py” file.

Figure 14.

IMG file backup process Create a backup of the OS image using “Win32 Disk Imager” (SourceForge)

Figure 15.

Website customization process

(A) Log in using the default account.

(B) Add a user account under the “Admin” tab.

(C) Add commands to be sent to RCMs (i), and edit or delete them as needed (ii).

Figure 16.

Device pairing process for global internet control

In the “Home” tab, scan for nearby devices around the mini-computer and selectively connect to target RCMs (i). Connected devices are automatically registered, and from the registered devices list, users can disconnect, reconnect, or deregister devices as needed (ii). Follow the steps in the indicated order (#1–#3).

Figure 17.

Real-time RCM control process

(A) In the “Send” tab, select RCMs and corresponding commands (i), then click “Send Selected Commands” (ii). After the commands are sent, confirmation messages appear below the button for 2–3 s.

(B) In the “Log” tab, view the history of sent commands (i), and optionally save it as a CSV file by clicking “Load CSV” (ii).

Figure 18.

Scheduled RCM control process In the “Scheduled Commands” tab, select one of the scheduling options – Schedule Once (i, top left), Schedule Periodic (i, bottom left), or Schedule Periodic (Advanced) (i, right) – and enter the detailed operation conditions

The configured commands are then created and listed based on the entered conditions (ii).

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and virus strains
pAAV-Syn-FLEX-rc[ChrimsonR-tdTomato]	Addgene	Cat# 62723-AAV5
pAAV_hSyn1-SIO-stGtACR2-FusionRed	Addgene	Cat# 105677-AAV1
Experimental models: Organisms/strains
Mouse: Agrptm1(cre)Lowl/J	The Jackson Laboratory	RRID:IMSR_JAX:012899; Male and female, 8–12 weeks
Mouse: B6.Cg-Gt(ROSA)26Sortm32(CAG-COP4∗H134R/EYFP)Hze/J	The Jackson Laboratory	RRID:IMSR_JAX:024109; Male and female, 8–12 weeks
Mouse: B6.Cg-Tg(Thy1-COP4/EYFP)18Gfng/J	The Jackson Laboratory	RRID:IMSR_JAX:007612; Male and female, 8–12 weeks
Mouse: B6.SJL-Slc6a3tm1.1(cre)Bkmn/J	The Jackson Laboratory	RRID:IMSR_JAX:006660; Male and female, 8–12 weeks
Software and algorithms
nRF Connect for Mobile	Nordic Semiconductor ASA	Download URL (Google Play Store); Download URL (Apple App Store)
Arduino IDE (version: 2.3.6)	Arduino	Download URL
SD Card Formatter (version: 5.0.3)	SD Association	Download URL
Win32 Disk Imager (version: 1.0.0)	SourceForge	Download URL
Ethovision XT (version: 13)	Noldus Information Technology BV	RRID:SCR_000441; Download URL
GraphPad Prism (version:10)	GraphPad	RRID:SCR_002798; Download URL
Other
Wireless optogenetic probe PCB	PCBWay	N/A (custom-made)
Solder paste	Chip Quik	SMDLTLFP10T5
Blue μ-ILED (470 nm)	Cree	TR2227
Orange μ-ILED (589 nm)	Three Five Materials	TCE10-589
Parylene C	Daisan Kasei Co., Ltd.	diX C
3-pin female connector	Harwin	M50-3130345
3D printer resin	B9Creations	BioRes – Medical/Wearable
Wireless control module PCB	PCBWay	N/A (custom-made)
BLE SoC	TAIYO YUDEN	EYSHSNZWZ
1 μF SMD capacitor (0603 metric)	Murata Electronics	GRM033C80G105ME05D
0.1 μF SMD capacitor (0603 metric)	Murata Electronics	GRM033R61A104KE15D
4.7 μF SMD capacitor (0603 metric)	Murata Electronics	GRM035R60J475ME15D
Red indicator LED (624 nm; 1005 metric)	VCC	VAOL-S4RP4
Green indicator LED (573 nm; 1005 metric)	VCC	VAOL-S4GT4
6.04 kΩ SMD resistor (0603 metric)	KOA Speer	RK73H1HTTC6041F
3.3 V low-dropout regulator	Onsemi	NCP4624DMU30TCG
3-pin male connector	Harwin	M50-3630342
2-pin male connector	Harwin	M50-3630242
2-pin female connector	Harwin	M50-3130245
25 mAh LiPo battery	Guangzhou Markyn Battery Co., Ltd.	GMB-400909
1100 mAh AAA battery	EBL	EBL AAA Rechargeable Ni-MH Batteries 1100mAh
Battery pack	ZRM&E	ZR-BC-9B26AS13A
M-M jumper wire	Gersangin	GSH-05401
BNC to alligator clip cable	Digilent	BNC to Alligator Clip Cable
6-pin female connector	Connfly	DS1023-1∗6S21
Atmospheric sensor module	SparkFun	SparkFun Atmospheric Sensor Breakout - BME280
Microcontroller board	Arduino	Arduino UNO Rev3
BLE module	OEM	HM-10 BLE Bluetooth V4.0 Module
64GB microSD card	SanDisk	SanDisk Extreme microSDXCTM UHS-I CARD - 64GB
Mini-computer	Raspberry Pi	Raspberry Pi 4 Model B
BLE dongle	Silicon Labs	BLED112
Stereotaxic alignment system, KOPF Model 942	KOPF	Model 942
Leica S9i Digital Stereo Microscope	Leica	Lecia S9i Digital Stereo Microscope
Charge-Coupled Device (CCD) camera	Any	Any

Step-by-step method details

Stereotaxic viral injection for optogenetics (social interaction experiment)

Timing: 1 h (5 weeks prior to implantation)

For step-by-step description of the stereotaxic surgery and viral injection procedure, please see “Part 1: Viral injection” in STAR Protocols by Tokizane and Imai (2025).²⁸ Here we describe the steps for bilateral injection of viral cocktail of AAV5-Syn-FLEX-ChrimsonR-tdTomato and AAV1-hSyn1-SIO-stGtACR2-FusionRed into the VTA of DAT-cre mice (Figures 19 and 20C).

• 1.Prepare the surgical area and stereotaxic frame for aseptic surgery, including sterilizing all surgical tools and instruments.
• 2.Anesthetize the mouse using 3% isoflurane and maintain an anesthetic plane using 1%–3% isoflurane.

CRITICAL: To ensure safe surgical procedure, monitor the animal’s respiration throughout the procedure and adjust isoflurane flow rate according to standard procedure.

• 3.
  Prepare the scalp for incision.

  • a.
    Shave the scalp and fix the animal’s head to the stereotaxic frame using the ear bars.
  • b.
    Apply eye lubricant and clean the incision site with betadine and 70% isopropyl alcohol with alternating application.
  • c.
    Apply topical lidocaine to the incision site.
• 4.
  Incise the skin along the center of the scalp from the anterior to posterior surface to expose the skull.

  • a.
    Clean the skull surface using 3% hydrogen peroxide solution and cotton-tipped swabs.
  • b.
    Ensure medial and lateral sutures and bregma and lambda are visible (Figure 19A).
• 5.
  Perform craniotomy.

  • a.
    Attach the drill to the stereotaxic frame. Align the drill bit to bregma and lambda, respectively, and use the ear bars and nose bar to level the skull if needed (Figure 19B).
  • b.
    Drill bilateral holes through the skull to fit a Hamilton Neuros Syringe or similar device at the VTA (−3.4 mm (AP); ±0.5 mm (ML); −4.5mm (DV) (Figure 19C).
• 6.
  Perform the microinjection.

  • a.
    Place the syringe into the holder/pump and attach the holder to the stereotaxic frame.
  • b.
    Align the needle tip to bregma and ensure proper stereotaxic measurement.
  • c.
    Align and lower the needle intracranially to the VTA (−3.4 mm (AP); +0.5 mm (ML); −4.5mm (DV).
  • d.
    Inject 175 nL viral cocktail of AAV5-Syn-FLEX-Chrimson-tdTomato and AAV1-hSyn1-SIO-stGtACR2-FusionRed (Figure 19D).
  • e.
    Repeat injection procedure to contralateral side for complete bilateral infusion.
  • f.
    Raise and remove injection needle.
• 7.Use suture to enclose incision site.
• 8.Allow animals to recover for 5 weeks following injection, then prepare for optogenetic probe implantation.

Figure 19.

Stereotaxic microinjection and implantation of probe for optogenetic experimentation

(A) Incision and exposure of skull to reveal bregma and lambda sutures.

(B) Alignment of drill bit to bregma and lambda sutures.

(C) Craniotomy to fit injection device and optogenetic probe.

(D) Microinjection of viral construct into the area of interest.

(E) Attachment and alignment of the optogenetic probe.

(F) Intracranial insertion of probe directed to the area of interest.

(G) Application of dental cement to secure probe to skull.

(H) Complete dental cement application.

(I) Connection to battery pack for behavioral experimentation.

Figure 20.

Schematic cartoon of surgical and behavior procedures

(A) Schematic of optogenetic probe implantation into the secondary motor cortex (M2) and behavioral apparatus for Locomotor Assay Experiment.

(B) Schematic of optogenetic probe implantation into the paraventricular nucleus of the hypothalamus (PVN) and behavioral apparatus for Food Intake Experiment.

(C) Schematic of microinjection/optogenetic probe implantation into the ventral tegmental area (VTA) and behavioral apparatus for Social Interaction Experiment.

Stereotaxic probe implantation for optogenetics (social interaction experiment)

Timing: 1 h (1 week prior to behavioral testing)

Here we describe the steps for implantation of bilateral, wireless μ-ILED optogenetic probes directly caudal to the VTA with the light path aimed rostrally into the structure (Figures 19 and 20A).

Note: Differences in targeting across brain regions are accommodated by selecting probe geometries and μ-ILED placement appropriate for the specific experimental application. In practice, probe length, μ-ILED positioning, and unilateral/bilateral configurations can be adjusted during device design to match the anatomical depth and spatial extent of the targeted structure. The experiments described illustrate how the same platform can be customized for different brain regions.

• 9.Anesthetize the mouse and prepare for intracranial surgery as described in Steps 1–5 above.
• 10.Drill holes through the skull to fit the probe at the VTA (−3.4 mm (AP); +0.5 mm (ML)).
• 11.
  Implant the probe.

  • a.
    Place the bilateral optogenetic probe into the cannula holder adapter and attach the adaptor to the stereotaxic frame.
  • b.
    Adjust the alignment of the probe in the adaptor to ensure the angle and direction of the probe are appropriate for proper illumination of the brain region of interest (Figure 19E).
  • c.
    Lower the bilateral probe intracranially to −3.4 mm (AP); +0.5 mm (ML); −4.5 mm (DV) with μ-ILEDs directed anteriorly towards the VTA (Figure 19F).
  • d.
    Secure and affix implant with dental cement, ensuring complete coverage of the skull (Figures 19G and 19H).
• 12.Allow the animal to recover in a clean, polycarbonate cage for 1 week before any behavioral testing.

Stereotaxic surgery and optogenetic probe implantation (locomotor assay experiment)

Timing: 1 h (1 week prior to behavioral testing)

Please follow the step-by-step procedure described above for implantation of bilateral, wireless μ-ILED optogenetic probes for optical stimulation of the secondary motor cortex (M2) in Thy1^ChR2-YFP mice (Figures 19 and 20A).

• 13.Anesthetize the mouse and prepare for intracranial surgery as described in Steps 1–5 above.
• 14.Drill holes through the skull to fit the bilateral probe at the motor cortex (+1.0 mm (AP); ±0.5 mm (ML)).
• 15.
  Implant the probe.

  • a.
    Place the optogenetic probe into the cannula holder adapter and attach the adaptor to the stereotaxic frame.
  • b.
    Adjust the alignment of the probe in the adaptor to ensure the angle and direction of the μ-ILED is directed medially towards M2.
  • c.
    Lower the probe intracranially to the motor cortex (+1.0 mm (AP); ±0.5 mm (ML); −0.5 mm (DV)) with the bilateral μ-ILED directed anteriorly towards M1.
• 16.Secure and affix implant with dental cement, ensuring complete coverage of the skull.
• 17.Allow the animal to recover in a clean, polycarbonate cage for 1 week before any behavioral testing.

Stereotaxic surgery and optogenetic probe implantation (food intake experiment)

Timing: 1 h (1 week prior to behavioral testing)

Here we describe the steps for implantation of unilateral, wireless μ-ILED optogenetic probes for optical stimulation of the paraventricular nucleus of the hypothalamus (PVN) in AgRP^Cre x Ai32 mice (Figures 19 and 20B).

• 18.Anesthetize the mouse and prepare for intracranial surgery as described in Steps 1–5 above.
• 19.Drill a hole through the skull to fit the probe at the PVN (−0.82 mm (AP); −0.5 mm (ML)).
• 20.
  Implant the probe.

  • a.
    Place the optogenetic probe into the cannula holder adapter and attach the adaptor to the stereotaxic frame.
  • b.
    Adjust the alignment of the probe in the adaptor to ensure the angle and direction of the μ-ILED is directed medially towards the PVN.
  • c.
    Lower the probe intracranially to (−0.82 mm (AP); −0.5 mm (ML); −5.25 mm (DV)) towards the PVN.
  • d.
    Secure and affix implant with dental cement, ensuring complete coverage of the skull.
• 21.Allow the animal to recover in a clean, polycarbonate cage for 1 week before any behavioral testing.

WNBN control for locomotion assay

Timing: 30 min

Here we describe the steps for controlling wireless optogenetic stimulation of Thy1-expressing neurons in the secondary motor cortex (M2) during the open field behavioral test (Figure 20A).

• 22.Move implanted Thy1^ChR2-YFP mice to the behavioral testing room and allow mice to acclimate undisturbed for 30 min.
• 23.
  Prepare and initiate locomotor activity test.

  • a.
    Set CCD camera and Ethovision software to monitor and record locomotor activity for 20 min.
  • b.
    Remove the experimental animal from the home cage and connect the RCM and battery pack to the intracranial probe.

    • i.
      Gently restrain the mouse by scruffing the back of the neck and firmly attach the RCM and battery pack connectors to the device.
  • c.
    Place animal into the square open-field enclosure (50 × 50 cm²) and initiate experiment start and video recording.
  • d.
    Use web interface to schedule stimulation commands to follow 10 min of baseline activity measurements.

    • i.
      Set the Bluetooth signal-initiated command to control a 20 Hz stimulation.
    • ii.
      Schedule the commands to be sent every minute to turn ON and OFF the μ-ILED, respectively, for 10 min.
• 24.Remove the mouse and return to home cage following 20 min of activity measurements.

Note: Ensure the RCM and battery packs are fully charged and operational prior to any behavioral testing.

WNBN control for food intake assay

Timing: 4 h 30 min

Here we describe the steps for controlling wireless optogenetic stimulation of AgRP-expressing neurons in the paraventricular nucleus of the hypothalamus (PVN) to drive food consumption in a food intake assay (Figure 20B).

• 25.Move implanted AgRP^Cre-ChR2 mice to the behavioral testing room.
• 26.
  Prepare and initiate food intake assay.

  • a.
    Set a large polycarbonate cage into the center of a sound attenuated box.
  • b.
    Set CCD camera and Ethovision software to monitor and record activity for 4 hs.
  • c.
    Remove the experimental animal from their home cage and connect the RCM and battery pack to the intracranial probe.
  • d.
    Gently restrain the mouse by scruffing the back of the neck and firmly attach the RCM and battery pack connectors to the device.
  • e.
    Place animal into a large polycarbonate cage for 1 h of acclimation and initiate experiment start and video recording.
  • f.
    Use web interface to schedule stimulation commands.

    • i.
      Set the Bluetooth signal-initiated command to control a 20 Hz stimulation.
    • ii.
      Schedule the ON command to be sent 2 h after experiment start to turn on μ-ILED. Schedule OFF command to be sent 3 h after experiment start.
• 27.After 1 h of acclimation, place a weigh boat containing one pellet of food (∼3 g) in center of cage for 3 h.
• 28.Remove and weigh food every h for the duration of the experiment.
• 29.Remove the mouse and return to home cage following 4 h experimental measurements.

Note: Ensure the RCM and battery packs are fully charged and operational prior to any behavioral testing.

WNBN control for social interaction assay

Timing: 10 min

Here we describe the steps for controlling wireless optogenetic stimulation and inhibition of VTA dopamine neurons in DAT-cre mice during a social interaction assay (Figure 20C).

• 30.Move mice (including age and sex matched “stranger” mice) to the behavioral testing room and allow mice to acclimate undisturbed for 30 min.
• 31.
  Remove the experimental animal from the home cage and connect the RCM and battery pack to their intracranial probe.

  • a.
    Gently restrain the mouse by scruffing the back of the neck and firmly attach the RCM and battery pack connectors to the device.
  • b.
    Place the mouse into a polycarbonate holding cage.
• 32.
  Prepare social interaction test.

  • a.
    Set the home cage into the center of a sound attenuated box.
  • b.
    Set a CCD camera and Ethovision software to monitor and record locomotor activity.
  • c.
    Use web interface to schedule stimulation (ON) command to be sent 2 min after experiment start to turn on μ-ILEDs. Schedule OFF command to be sent 4 min after experiment start.
  • d.
    Set the Bluetooth signal-initiated command to control a 20 Hz stimulation period during the introduction of the stranger mouse.

    • i.
      For optogenetic stimulation, set the command to red LEDs.
    • ii.
      For optogenetic inhibition, set the command to blue LEDs.
    • iii.
      Counterbalance groups for order of light stimulation.
• 33.
  Begin social interaction test.

  • a.
    Place the test animal into the home cage and allow it to explore for 1 min.
  • b.
    Following 1 min, place the stranger mouse into the home cage.
  • c.
    Allow the mice to interact for the remainder of the test (4 min).
  • d.
    Remove the mice and return to appropriate cages.
• 34.
  Perform behavioral testing and data analysis.

  • a.
    Define and manually score any period of time in which the test mouse interacts with the stranger mouse. Interaction includes sniffing the stranger’s snout, flank, or anogenital area, grooming, or following the stranger as it explores the cage.
  • b.
    Blind all behavioral data to the experimenter prior to manual scoring.

Note: The above remote, high-throughput experiments can also be conducted using traditional in vivo behavioral laser equipment.

WNBN control for remote, automated data collection

Timing: 2 days

Here we describe the steps for automated data collection of home cage temperature and air pressure using sensor modules.

• 35.
  Install the sensor modules on the mouse cage.

  • a.
    Place Arduino board on top of the cage and insert sensor module though cage lid.
  • b.
    Ensure sensor module is not accessible to animals.
  • c.
    Power Arduino board via USB or power connections.
• 36.Access the internet website and set scheduled commands.

Note: The location of experimenter can be anywhere, regardless of the location of the home cage where the sensor is installed.

• 37.Examine collected data either on the ‘Log’ page or ‘Received Data’ page during or after monitoring.

Expected outcomes

The behavioral experiments described above are based on classical optogenetic behavioral experiment procedures with well-established, published outcomes.^{1,29,30,31,32,33} These behavioral experiments were selected for their robust, repeatable behavioral effects. For the control for locomotion assay, we expect that ChR2 stimulation of secondary motor cortex will increase locomotion in Thy1^ChR2-YFP mice, including increased total distance traveled and total number of rotations compared to no stimulation.³² For the food intake assay, we expect that AgRP^Cre x Ai32 mice will consume significantly more food during 20 Hz photostimulation (1 h) compared to non-stimulation periods.^29,30,31 While the effect size might be altered between approaches due to light power and targeting strategies, we expect the same general outcomes for both wireless stimulation and stimulation with conventional fiber optics.¹ For the social interaction assay, we expect that Chrimson stimulation (20 Hz) will significantly increase social interaction for DAT^{Cre-Chrimson/GtACR2} mice. In contrast, we expect that 40 Hz stimulation stGtACR2-mediated neuronal inhibition will reduce social interaction in DAT^{Cre-Chrimson/GtACR2} mice.³³ For the control of remote, automated data collection, we expect that automated data collection of home cage temperature and air pressure for home cages will successfully record temperature changes in cages that contain mice as compared to empty cages. We also anticipate an increase in temperature during the dark cycle as compared to the light cycle. These changes are likely due to increased activity of the nocturnal mice during the dark cycle. If adopters of this protocol deploy the WNBN in novel neuroscience experiments, the outcomes of those will depend on experimental details, but users can expect successful remote control of optogenetic constructs as observed by immediate early gene transcription, electrophysiology, and/or behavior.

Quantification and statistical analysis

Animal behavioral data including distance traveled, time spent moving, and social interaction time are collected via CCD camera and calculated using Ethovision 13 (Noldus Information Technologies, Leesburg, VA). Food intake measurements are manually weighed. All data are analyzed and visually represented using graphic and statistical software GraphPad Prism 10. Data are analyzed for statistical significance by either repeated measures one-way ANOVA with Tukey’s multiple-comparisons correction or paired, two-tailed t test. Parametric tests were used when data were normally distributed.

Limitations

Despite the modularity and scalability of the WNBN platform for wireless control of multiple RCMs in neuroscience applications, several practical limitations affect its performance and applicability.

First, in local piconet control mode, the number of simultaneously connected RCMs is limited to approximately 15 due to inherent constraints in smartphone operating systems and BLE stack implementations, which typically only support a limited number of concurrent BLE connections. In contrast, global internet control mode – where a Raspberry Pi functions as a centralized controller managing multiple BLE dongles – offers greater scalability. While the system has been empirically validated with up to 21 devices, the actual upper limit depends on factors such as the number of BLE dongles connected via USB, the Raspberry Pi’s capacity to handle multiple Bluetooth stacks, CPU load, and the efficiency of the connection scheduling algorithm. As more RCMs are added, communication latency may increase and issues such as packet collisions or scheduling delays can arise due to the non-deterministic nature of BLE advertisement scanning and connection intervals. Such complications are exacerbated when the existing BLE environment is crowded with devices not associated with the experiments at hand. For this reason, we strongly encourage removing unnecessary BLE devices from the laboratory during active deployment of the WNBN. Simple changes such as switching to wired keyboards and computer mouses and removal of unrelated smartphones and smartwatches can greatly reduce local BLE traffic. These issues could potentially be mitigated by adopting a Bluetooth Mesh architecture, which enables decentralized communication through message relaying across nodes; however, this would require substantial modifications to the firmware and control protocol.

Second, the above issues of scale and BLE crowding have important follow-on implications for the type of experiment used with WNBN. While piconet, local internet, and global internet applications all have relatively short latencies (∼25–150 ms depending on modality), interrupted signals can have varying degrees of impact on experimental conditions if multiple attempts are needed to send or receive the BLE-based commands. In an hour-long continuous photostimulation such as with the feeding experiments described here, the impact of even a few seconds long delay would be minimal, but in settings where more tightly-controlled photostimulation is necessary (e.g., closed-loop operant self-stimulation) then we recommend maintaining local control, reducing extraneous BLE traffic, and carefully monitoring error logs for missed signals and delays. As a best practice, any experiment that is dependent on low latency signal transmission should report these values along with experimental outcomes.

Third, the current BLE-based implementation imposes limitations on data packet size and sampling rate, which pose challenges for high-throughput neural recording or time-sensitive acquisition of other biological signals that require greater bandwidth and finer temporal resolution. While the platform is well-suited for wireless control tasks such as optogenetic stimulation, it is less appropriate for applications involving continuous streaming of complex physiological data. Future iterations of the system may consider integrating alternative wireless protocols – such as Wi-Fi or ultra-wideband – to support these use cases, but each potential change carries important trade-offs in power consumption, protocol complexity, and system size.

Fourth, BLE communication relies on strict matching of device identifiers, particularly the UUIDs assigned to read and write characteristics. If the firmware-defined UUIDs do not match those expected by the mobile app or web-based control software, pairing failures or command transmission errors may occur. This issue is especially relevant during firmware updates or when new hardware versions are introduced. For example, switching from an optogenetic module to an HM-10-based environmental sensor without updating the UUID mappings in the control software may result in silent pairing failures or unstable communication. Users must therefore ensure consistency of UUID settings across both the software and firmware layers to maintain reliable operation.

Finally, the wireless methods used here are battery-based, providing inherent limitations to the duration of possible experiments. The experiments using wireless devices are dependent on the stimulation modality as well as the duration of activation of each modality. For example, short duration pulse trains will last much longer than constant illumination of μ-ILEDs. We also note than even the method for pairing with conventional hardware presented here uses a rechargeable, consumer-grade battery-based approach. While the lifetime from these AAA batteries is more than sufficient for typical experiments, these batteries might need to be recharged and/or replaced during a very long duration experiment. In any case, users should derive estimates of battery lifetimes prior to beginning the full WNBN-based experiments.

Troubleshooting

Problem 1

Connection failure or control software crash during BLE pairing (related to – Setup of RCC for local piconet control, Step 12; Setup of RCC for global internet control, Step 16).

Even when the RCM appears in the BLE scan list, connection attempts may fail or cause the app or web interface to crash. This typically occurs when the UUIDs embedded in the BLE firmware do not match those expected by the control interface. BLE scanning itself does not validate service or characteristic UUIDs, so mismatches are not detected until the connection is initialized and the GATT table is queried. If the expected UUID is not found at this stage, the connection attempt may silently fail or trigger abnormal behavior in the application. This issue often arises when switching between different firmware types (e.g., optogenetic vs. HM-10 sensor modules) without updating the UUID setting on both the firmware and software sides.

Potential solution

Confirm that the UUIDs for the write/read characteristics in the BLE firmware match exactly with the UUIDs defined in the control application. For optogenetic RCMs, the required UUID is f3641401-00b0-4240-ba50-05ca45bf8abc; for sensor modules based on the HM-10 module, it is 0000ffe1-0000-1000-8000-00805f9b34f. A mismatch in these values will not be flagged during scanning, but will prevent pairing or cause the control software to crash after partial connection.

In global internet control mode using Raspberry Pi, the UUID must also be correctly set in the “controller.py” script, along with the proper BLE address type (‘random’ for SoCs like EYSHSNZWZ, and ‘public’ for fixed-MAC modules like HM-10). If the device was previously paired with a different configuration or app version, cached BLE sessions may interfere – always unpair, reboot, and rescan before retrying connection. Re-upload the appropriate firmware if inconsistencies are suspected, and avoid mixing modules of different UUID schemes in a single control session unless all UUIDs are explicitly handled in the backend.

Problem 2

Failure to load the global web interface or encountering “This site can’t be reached” or “404 Not Found” errors (related to – Setup of RCC for global internet control, Step 15).

This issue typically occurs when the microSD card was not flashed correctly, or the necessary peripherals (BLE dongle, LAN cable, etc.) were not properly inserted prior to powering on the mini-computer. If the device is not connected to the network, or if the server fails to boot properly due to missing components, a browser attempting to access the web interface will show the error “This site can’t be reached.” In contrast, when the mini-computer is successfully connected to the network but the backend service is not yet initialized, the browser will show a “404 Not Found” message. These two errors reflect different failure modes and should be addressed separately.

Potential solution

For the “This site can’t be reached” error, verify that the microSD card has been correctly flashed with the appropriate IMG file and that all required peripherals – including the BLE dongle and LAN cable – are inserted before applying power. The power cable must be connected last to ensure proper boot sequence. After startup, open the terminal and run ‘ping 8.8.8.8’ to check for internet connectivity. A successful connection will return “64 bytes from 8.8.8.8…” If this response does not appear, the device may not have received an IP address or may be blocked by firewall settings. Contact your institutional IT department to confirm whether MAC address registration is needed.

For the “404 Not Found” error, wait at least 5 min after booting to allow the backend services to fully initialize before accessing the interface. If the problem persists, try clearing the browser cache and reloading the page, as a cached error may persist even after the server is ready. If the issue is still unresolved, the backend service may have failed to start due to a corrupted image or improper shutdown. As a last resort, reflash the microSD card using the previously backed-up IMG file (Step 14c) by repeating Step 13a–b.

Problem 3

Increased latency or command failures during global internet control with multiple RCMs (related to – Setup of RCC for global internet control, Steps 16 and 17).

While latency in local piconet mode is typically low (∼20–50 ms) and reliable for up to ∼15 devices, global internet control introduces higher latency and greater variability depending on how many RCMs are assigned per BLE dongle. Although each BLE dongle can theoretically support up to eight connections, empirical tests show that stability declines when more than 3–4 RCMs are connected to a single dongle. Beyond this point, issues such as packet loss and delayed command execution are common, due to bandwidth limitations and the sequential nature of BLE communication scheduling. These limitations can significantly impact system responsiveness and cause failure of scheduled commands during behavioral experiments.

Potential solution

To enhance reliability and reduce latency, users can distribute RCMs across more BLE dongles. Although the Raspberry Pi has only four native USB ports, this limitation can be overcome using a powered USB hub. One-to-one dongle-to-RCM mapping is not required; rather, balancing the number of RCMs across dongles helps optimize performance and reduces per-dongle communication load.

If command failures persist repeatedly for a specific RCM, it is recommended to manually disconnect and deregister the device via the “Registered Chips” list in the web interface, then re-register it by scanning and pairing again before resending the intended command. This process helps reset the session and may resolve transient pairing or communication issues for that device.

Problem 4

Loss of wireless battery pack/module connection during behavioral testing (related to – behavioral execution steps; not limited to a single step, including WNBN control for locomotion assay, food intake assay, social interaction assay, and remote, automated data collection).

Potential solution

Confirm a secure connection between the battery pack and the implant prior to testing. If the connection is not secure, select an alternative battery pack that will ensure a firm connection. Alternatively, if the plug connection is not tight following implantation, it is possible apply a small amount adhesive tape to secure the pack to the probe.

Problem 5

Loss of battery power/connectivity during behavioral testing (related to – behavioral execution steps; not limited to a single step, including WNBN control for locomotion assay, food intake assay, social interaction assay, and remote, automated data collection).

Potential solution

Confirm all battery packs are fully charged and functioning properly prior to testing. On a benchtop, using a separate optogenetic probe, connect each battery pack and send stimulation parameters to confirm μ-ILED activation. When using a 25 mAh LiPo battery (GMB-400909, Guangzhou Markyn Battery Co., Ltd.) with the wireless control module and optogenetic probe, the system typically operates for approximately 6.5–10 h when the μ-ILED is driven continuously, depending on stimulation parameters (e.g., stimulation frequency such as 5, 10, 20, or 40 Hz with 10 ms pulse width). If the μ-ILED is activated intermittently or stimulation is paused for extended periods, battery life may be extended. If the expected duration of the experiment approaches or exceeds the battery capacity, consider reducing stimulation parameters or shortening the experiment duration.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jordan G. McCall (jordangmccall@wustl.edu).

Technical contact

Technical questions on executing this protocol should be directed to and will be answered by the technical contact, Jae-Woong Jeong (jjeong1@kaist.ac.kr).

Materials availability

All custom hardware components used in this protocol—including Gerber files for the wireless optogenetic probe and wireless control module PCBs, 3D CAD file for the probe holder, and the IMG files for the mini-computer OS—have been deposited at Zenodo: https://zenodo.org/records/15388068. This study did not generate new unique reagents.

Data and code availability

The published article includes all firmware files, Arduino sketches, and Python scripts necessary for operating the WNBN platform. These include over-the-air firmware update packages, source code for BLE control, and Arduino code for the sensor modules. All files have been deposited at Zenodo: https://zenodo.org/records/15388068.

Acknowledgments

This work was supported by the National Research Foundation of Korea (RS-2024-00335066, RS-2022-NR068144, and RS-2025-02218624 to J.-W.J. and RS-2025-00556794 to C.Y.K.). The authors also acknowledge support from the National Institutes of Health (R01NS117899 and R21DA055047 to J.G.M.). The graphical abstract and Figure 20 were created in part using BioRender.com.

Author contributions

Conceptualization, C.Y.K., K.E.P., J.-W.J., and J.G.M.; probe and RCM hardware design, C.Y.K. and R.Q.; probe and RCM hardware fabrication, C.Y.K., D.H., Y.J., S.-Y.J., and C.C.; firmware development for RCM hardware, C.Y.K. and R.Q.; software development for local piconet control and global internet control, C.Y.K., R.Q., J.C., S.H., and J.-W.J.; in vivo experimental design, K.E.P. and J.G.M.; surgeries, K.E.P.; writing, C.Y.K., K.E.P., J.-W.J., and J.G.M.; figure preparation, C.Y.K., K.E.P., E.Y.J., D.H., and J.L.; funding acquisition, C.Y.K., J.-W.J., and J.G.M.; supervision, J.-W.J. and J.G.M.

Declaration of interests

The authors declare no competing interests.