A low-cost electroencephalographic recording system combined with a millimeter-sized coil is proposed to drive transcranial magnetic stimulation of the mouse brain in vivo. Using conventional screw electrodes with a custom-made, flexible, multielectrode array substrate, multi-site recording can be carried out from the mouse brain in response to transcranial magnetic stimulation.
A low-cost electroencephalographic (EEG) recording system is proposed here to drive transcranial magnetic stimulation (TMS) of the mouse brain in vivo, utilizing a millimeter-sized coil. Using conventional screw electrodes combined with a custom-made, flexible, multielectrode array substrate, multi-site recording can be carried out from the mouse brain. In addition, we explain how a millimeter-sized coil is produced using low-cost equipment usually found in laboratories. Practical procedures for fabricating the flexible multielectrode array substrate and the surgical implantation technique for screw electrodes are also presented, which are necessary to produce low-noise EEG signals. Although the methodology is useful for recording from the brain of any small animal, the present report focuses on electrode implementation in an anesthetized mouse skull. Furthermore, this method can be easily extended to an awake small animal that is connected with tethered cables via a common adapter and fixed with a TMS device to the head during recording.The present version of the EEG-TMS system, which can include a maximum of 32 EEG channels (a device with 16 channels is presented as an example with fewer channels) and one TMS channel device, is described. Additionally, typical results obtained by the application of the EEG-TMS system to anesthetized mice are briefly reported.
Transcranial magnetic stimulation (TMS) is a promising tool for human brain science, clinical application, and animal model research because of its non-/low invasiveness. During the early stage of TMS applications, measurement of the cortical effect in response to single- and paired-pulse TMS in humans and animals was restricted to the motor cortex; easily measurable output was limited to motor evoked potentials and induced myoelectric potentials involving the motor cortex1,2. To expand the brain regions that can be measured by TMS modulation, electroencephalographic (EEG) recording was integrated with single- and paired-pulse TMS as a useful method to directly examine the excitability, connectivity, and spatiotemporal dynamics of areas throughout the whole brain3,4,5. Thus, the simultaneous application of TMS and EEG recording (TMS-EEG) to the brain has been used to probe various superficial cortical brain areas of humans and animals to investigate intracortical neural circuits (see Tremblay et al.6). Moreover, TMS-EEG systems can be used to examine additional cortical spatiotemporal characteristics, including the propagation of signals to other cortical areas and the generation of oscillatory activity7,8.
However, the mechanism of action of TMS in the brain remains speculative because of the non-invasiveness of TMS, which limits our knowledge of how the brain functions during TMS applications. Therefore, invasive translational studies in animals ranging from rodents to humans are of crucial importance to understand the mechanism of the effects of TMS on neural circuits and their activity. In particular, for combined TMS-EEG experiments in animals, a simultaneous stimulation and measurement system has not been intensively developed for small animals. Therefore, experimentalists are required to construct such a system by trial and error according to their specific experimental requirements. In addition, mouse models are useful among other in vivo animal species models because many transgenic and strain-isolated mice strains are available as biological resources. Thus, a convenient method to build a TMS-EEG-combined measurement system for mice would be desirable for many neuroscience researchers.
This study proposes a TMS-EEG-combined method that can be applied for simultaneous stimulation and recording of the mouse brain, which is the main type of transgenic animal used in research, and that can easily be constructed in typical neuroscience laboratories. First, a low-cost EEG recording system is described using conventional screw electrodes and a flexible substrate to reproducibly assign an electrode-array position in each experiment. Second, a magnetic stimulation system is constructed using a millimeter-sized coil, which can easily be custom-made in typical laboratories. Third, the TMS-EEG-combined system records neural activity in response to sound and magnetic stimulation. The method presented in this study can reveal the mechanisms that generate specific disorders in small animals, and the results obtained in the animal models can be translated to understand the corresponding human disorders.
In the present study, all animal experiments were performed following the National Institutes of Health Guide for the Care and Use of Laboratory Animals and with approval from the Institutional Animal Care and Use Committee of Hokkaido University. C57BL/6J mice, two male and three female, 8 to 10 weeks old, were used for the present study. This is a terminal procedure. The animals were obtained from a commercial source (see Table of Materials).
1. Flexible two-dimensional array design and construction
Figure 1: Component parts of the flexible two-dimensional (2D) array for electroencephalographic (EEG) recording and the fabricated device including the array. (A) The miniature screw electrode that is embedded in the mouse skull. (B) The designed electrode pads for measuring brain activity (green circles) and the reference channel (square on the bottom right). The relative coordinates of the electrode pads from a reference point (cross mark) at the origin (0, 0) are shown; the size in millimeters is illustrated in parentheses. The center coordinates of the electrode pads are symmetrical with respect to the vertical axis passing through the cross mark. (C) The electrode pads and drill holes for a recording electrode (left) and a reference electrode (right) are illustrated. (D) A surface-mount connector (2 × 10 pins) used for the flexible 2D array (left) and the pattern and size of the designed pads on the substrate (right). (E) Designed blueprint with the size of each part in millimeters. (F) Image of a fabricated substrate indicated by the blueprint in E. (G) The layer structure of the flexible 2D array (head and connector parts). The top and side views of the screw electrode pads (top) and readout pads (bottom) are illustrated. The head and the connector parts are composed of a three-layered structure (top) and a six-layered structure (bottom), respectively. Additionally, the neck part is composed of a five-layered structure; a protective polyimide layer is mounted on the top and back surface, and the reinforcing polyimide board is not mounted on the neck part. Please click here to view a larger version of this figure.
2. Adaptor construction and channel mapping
Figure 2: Constructing the adaptor for a two-dimensional (2D) electrode array on the flexible substrate and recording channel mapping. (A) In the connector part, the reference and ground channels are connected to the bottom electrode pads with lead wires. If the reference and ground channels are determined in advance, the channels should be connected to the corresponding bottom electrode pads during the design phase. In such cases, soldering lead wires to the channels and electrode pads is unnecessary. (B) Insulation-displacement connectors (top left) are crimped to one end of the flat cable (bottom left) to link the measurement amplifier connector (top right). All lines that correspond to the channels to be used are soldered to the green connectors (bottom right). In this case, because each green connector connected to the head amplifier is assigned for an eight-channel measurement, at least two connectors are needed to record 16-channel brain activity signals.The soldered points are covered with epoxy resin and shielding tape to prevent contact with other signal lines. (C) The connector and fabricated cable are placed on the surface of the flexible 2D array substrate. The thin stainless-steel rod is attached to the back side of the flexible substrate. (D) The spatial locations of recording channels on the mouse brain surface and the channel maps for each point for the measurement system are shown. In this case, there are 16 recording channels with screw electrodes (red circles), although the total number of possible recording sites is 32. The other 16 non-recording channels are also shown as green circles on the brain surface. In the mapping plot, "G" and "R" indicate the channels designed for ground and reference electrodes, respectively. Please click here to view a larger version of this figure.
3. Animal surgery
4. Electrode implantation
5. Small coil design and construction
Figure 3: Small coil for magnetic stimulation. (A) Three-dimensional (3D)-printed disk (left). Two identical disks are adhered to the permalloy-45 shaft; one is at the end of the shaft, and the other is 10 mm away (right). (B) Setup for winding the coil. The 60 mm shaft with the two disks is attached to an impact driver. A hall-effect sensor is placed near the small magnet attached to the shaft. The copper wire is wound between the two disks. (C) Constructed coil. The coil is 10 mm in height, 6 mm in diameter, and has 1,000 turns of copper wire. The right side of the figure shows the coil manipulated by a 3D-printed coil holder. (D) AC properties of the coil recorded by an LCR meter: (top) resistance versus frequency of sinusoidal input; (bottom) inductance versus input frequency. A typical coil has a resistance and inductance of 21.6 and 7.9 mH, respectively, at 1 kHz of AC input. (E) Biphasic rectangular waveform used as the coil input recorded by an oscilloscope. (F) Relationship between magnetic flux density and the distance between a constructed coil and the hall-effect sensor. The magnetic flux density was recorded by five different hall-effect sensors, once for each sensor. The average of five measurements is plotted, and error bars represent the standard errors of the mean. Please click here to view a larger version of this figure.
6. Signal recording system and procedure
7. Data analysis
Sample EEG data recorded in anesthetized C57BL/6J mice with the flexible substrate combined with the screw electrodes are presented below.
As a typical example, the average EEG waveforms generated in response to sound stimulation (8 kHz tone-burst, 80 dB sound pressure level [SPL]) are shown for 60 trials with identical stimuli (Figure 4A). A schematic of recording channel mapping is also presented in the middle of Figure 4A. The responses from Chs 5, 7, 10, and 12 are recorded from areas near the auditory cortex in both temporal lobes. In the individual EEG waveforms of the channels located around the auditory areas (the inferior colliculus and auditory cortex), the responses excluding the stimulation artifacts were first negative-going immediately after sound stimulation onset (e.g., Chs 3 and 10); the peak amplitudes were 45.6 ± 4.0 µV and25.6 ± 1.5 µV, respectively. The responses were subsequently positive-going to some extent over the baseline (Figure 4B,C) and oscillating while dampening. In contrast, responses from other channels were nearly independent of the stimulation onset, although some channel waveforms showed similar responses.
Figure 4: Sound event-related potential (ERP) waveforms at 16 sites in the mouse brain. (A) In response to sound (8 kHz tone-burst, 80 dB SPL) stimulation applied to an anesthetized mouse, 16-channel ERP waveforms are illustrated. The schematic of a mouse brain is shown in the center, and the 16 recording sites (red circles) on the mouse brain surface are indicated by channel numbers. In this case, 16 recording channels are used; the other 16 non-recording channels are shown as green circles. (B) Expanded views of ERP waveforms for Ch 3. (C) Expanded views of ERP waveforms for Ch 10. Please click here to view a larger version of this figure.
Similarly, the average waveforms of EEG recordings in response to short magnetic stimulation (Vin = 60 Vpp) of the area near the right inferior colliculus are shown for 60 trials with identical stimuli in Figure 5A. A schematic of recording channel mapping is also presented in the middle of Figure 5A. Because the stimulation coil was located near the area of Ch 14, the stimulation artifact was largest at that channel. However, relatively large stimulation artifacts were observed for most channels immediately after stimulation onset, indicating that the magnetic stimulation influenced all recording sites. Because the responses from Chs 5, 7, 10, and 12 were recorded from areas near the auditory cortex in both temporal lobes, the individual EEG waveforms excluding the stimulation artifacts were first negative-going and then positive-going to some extent, depending on the channel positions (Figure 5A–C). Near the auditory areas, response time courses induced by magnetic stimulation were different from those induced by sound stimulation. For Chs 3 and 10, for example, the responses were negative-going immediately after sound stimulation onset, although the peak amplitudes were 58.8 ± 4.0 µV and 28.2 ± 2.0 µV, respectively. Furthermore, with increasing magnetic stimulation intensities, the peak amplitudes of driven responses for Ch 10 were increased (Figure 5D), suggesting that the magnetic stimulation affected evoked neural responses.
Figure 5: Transcranial magnetic stimulation (TMS)-driven event-related potential (ERP) waveforms at 16 sites in the mouse brain. (A) The 16-channel ERP waveforms in response to TMS (Vin = 60 Vpp) applied to an anesthetized mouse are illustrated. A schematic of a mouse brain is shown in the center, and the 16 recording sites (red circles) on the mouse brain surface are indicated by the channel numbers. (B) Expanded views of ERP waveforms for Ch 3. (C) Expanded views of ERP waveforms for Ch 10. (D) Summary for the amplitudes of Ch. 10 ERPs evoked by different magnetic intensities (input voltage). For statistical analysis, an ANOVA for multiple comparisons followed by a post-hoc Tukey-Kramer test is used. * and *** represent p < 0.05 and p < 0.001, respectively. The trial number for a session is 60 times for each condition of individual animals. The statistics are calculated for the samples obtained from two animals. Please click here to view a larger version of this figure.
This method can also be easily extended to an awake small animal that is connected with tethered cables via a common adapter and fixed with a TMS device to the head during recording (Supplementary Figure 1 and Supplementary Figure 2).
Supplementary Figure 1: Fixture of the stimulation coil attached to a mouse skull. (A) For an awake mouse, a stimulation coil fixed with the fixture attached to the mouse skull is shown. (B) Event-related potentials (ERPs) of the awake mouse were recorded in an acryl box, where the mouse could move inside the box. Please click here to download this File.
Supplementary Figure 2: Waveforms of sound-driven and transcranial magnetic stimulation (TMS)-driven ERPs at 16 sites from the brain of an awake mouse. (A) In response to sound stimulation (8 kHz tone-burst, 80 dB SPL) applied to an awake mouse in an acryl case (Supplementary Figure 1B), 16-channel ERP waveforms are illustrated. The schematic of a mouse brain is shown in the center, and the 16 recording sites (red circles) on the mouse brain surface are indicatedby channel numbers. In this case, 16 recording channels are used; the other 16 non-recordingchannels are shown as green circles. (B) Similarly, 16-channel ERP waveforms in response to TMS (Vin = 60 Vpp) applied to the same awake mouse are illustrated. A schematic of a mouse brain is shown in the center, and the 16 recording sites (red circles) on the mouse brain surface are indicated by the channel numbers. The stimulation coil is located near the area of Ch 14. Please click here to download this File.
Supplementary Coding File 1: CAD data file for the donut-shaped disk required for the coil construction. Please click here to download this File.
This study addresses a multi-site EEG recording system combined with a magnetic stimulation system designed for small animals, including mice. The constructed system is low-cost and easily constructed in physiological laboratories, and can extend their existing measurement setups. The surgical procedure necessary to obtain data from the mouse recording system is profoundly simple if such laboratories have previous experience with standard electrophysiological experiments.
One advantage of using this approach is the good reproducibility of electrode placement on an individual animal's head and scalp. The flexible substrate used to assign screw electrodes to brain target sites is easily replicated using standard microfabrication techniques, and the same substrates are also convenient for determining the recording sites from the scalp of each animal. In addition, the shape of the electrode array can be easily modified to optimize various experimental needs; customized electrode arrangements can be optimally created for specific experimental purposes. If the method stated in the protocol is followed, screw electrodes, connectors, cables, and surgical procedures can be easily modified and extended to a measurement system with a larger number of recording sites. A second advantage of this recording system is its low cost when laboratories are equipped with a multichannel amplifier. The present recording system can obtain neural signals from 32 input channels and up to four separate cables. Therefore, an extended 32-channel recording system would require extra cables, screw electrodes, and modified flexible substrates, and this expanded system would have a very low cost.
However, one drawback of this methodology is the precise control of the depth of screw electrodes during implantation. However, this drawback is always present for typical screw EEG electrodes, and the precise depth of the screws premortem relative to the cortical surface is unknown. Furthermore, in this system, another critical point for the recording quality of EEG signals and reducing the noise level is appropriate electrode contact with the epidural layer. We always confirm appropriate electrode contact of all screw electrodes through impedance measurement. Typically, an impedance of 5-10 kΩ at 1 kHz suggests appropriate epidural placement, and the impedance values should be confirmed before neural signal measurement.
Additionally, in the current protocol, dental cement is applied to the skull before electrode implantation. The appropriate amount of dental cement can affect the success of EEG signal recordings. That is, a thin layer of dental cement on the skull does not support the implanted electrodes or fix the position of the electrodes, whereas a thicker layer prevents proper positioning of the electrode(s) to contact the dura matter. To determine the appropriate thickness of the layer, we measured the thickness of the dental cement using a digital caliper after successful EEG recordings. The average thickness of an appropriate cement layer was 0.7 mm, suggesting that the dental cement layer could be replaced with a "skull cap" with a 0.7 mm thickness and small holes for screw electrodes.
Magnetic stimulation is a useful tool in human and animal studies for minimally invasive or noninvasive neurostimulation of the brain. Rapidly changing the currents in a coil creates a magnetic field around the coil and causes hyperpolarization or depolarization of neuron membranes when the currents pass through animal and human skulls. For animal models, action potential responses are directly kindled by the suprathreshold magnitude of the electric field change, whereas subthreshold changes in neural membranes are produced to attune the network activity of neural populations10. This coil is simulated to produce an electric field that is more than 10 V/m, up a depth of 1.8 mm from the surface of the brain (2.4 mm from the skull), corresponding to cortical layer 5/6 or deeper regions in a typical (e.g., C57BL/6J) mouse10. These millimeter-sized coils are capable of inducing suprathreshold neural activity and can even generate a more localized electric field on the surface of the brain compared to those induced by previously reported coils11. Although added effects consisting of several factors, including perceived sound, skull vibration, and the thermal effect, cannot be completely excluded, these individual effects had a little influence on neural activity. Moreover, as a magnetic core, we use permalloy, whose magnetic properties usually depend on the conditions of the annealing process, including the cooling rate, annealing temperature, and holding time12. However, its annealing conditions could not be controlled as it was a commercial permalloy.
Recently, combined measurement systems consisting of multi-site EEG recording and TMS have been used in medical studies, and their clinical applications have been increasing4,6. Our proposed approach will improve small animal models (particularly mouse models) of human neurophysiology, which can provide a much easier translation of experimental rodent model results to human clinical counterparts by offering animal models that better parallel human systems. Finally, using multi-site recording techniques in genetically modified mice, combined magnetic and pharmacological interventions in animals with sensory hearing loss could help to reveal the mechanisms generating specific auditory disorders and tinnitus, which are our future research targets.
The authors have nothing to disclose.
This work was supported by the Murata Science Foundation, the Suzuken Memorial Foundation, the Nakatani Foundation for Advancement of Measuring Technologies in Biomedical Engineering, and a Grant-in-Aid for Exploratory Research (grant number 21K19755, Japan) and for Scientific Research (B) (grant number 23H03416, Japan) to T.T.
3D printer | Zhejiang Flashforge 3D Technology Co., Ltd | FFD-101 | The printer used for 3D-printing the donut-shaped disks |
ATROPINE SULFATE 0.5 mg | NIPRO ES PHARMA CO., LTD. | – | Atropine sulfate |
Bipolar amplifier | NF Corp. | KIT61380 | For amplifying waveforms for coil input |
Butorphanol | Meiji Seika Pharma Co., Ltd., Tokyo, Japan |
– | For anathesis of animals |
Commercial manufacturer of flexible 2D array | p-ban.com Corp. | – | URL: https://www.p-ban.com/ |
Computer prograom to analyze output signals | Natinal Instruments | NI-DAQ and NI-DAQmx Python | To analyze output signals from the hall-effect sensor |
Connector | Harwin Inc. | G125-FV12005L0P | For connector to conect to the measuring system |
Copper pad | p-ban.com Corp. | copper | Copper pad on each substrate |
Copper wire | Kyowa Harmonet Ltd. | P644432 | The windings of the coil |
DAQ board | National Instruments Corp. | USB-6343 | For measuring the magnitic flux density of the coil |
Dental cement | SHOFU INC. | Quick Resin | Self-Curing Orthodontic Resin |
ECoG electrode | NeuroNexus Inc. | HC32 | For reference to design of the flexible 2D array |
Epoxy resin | Konishi Co. Ltd. | #16123 | For coil construction |
Ethyl Carbamate | FUJIFILM Wako Pure Chemical Corp. | 050-05821 | For urethan anesthesia |
Flat ribbon cable | Oki Electric Cable Co., Ltd. | FLEX-B2(20)-7/0.1 20028 5m | For cable to connect between surface-mount connector and measuring sysytem |
flexible substrate | p-ban.com Corp. | polyimide | Baseplate of flexible substrate |
Function generator | NF Corp. | WF1947 | For generating waveforms for coil input |
Hall-effect sensor | Honeywell International Inc. | SS94A2D | For measuring the magnitic flux density of the coil |
IDC crimping tool | Pro'sKit Industries Co. | 6PK-214 | To crimp the IDC and one end of the flat ribbon cable; Flat cable connector crimping tool |
Instant glue | Konishi Co. Ltd. | #04612 | For coil construction |
Insulation-displacement connector (IDC ) | Uxcell Japan | B07GDDG3XG | 2 × 10 pins and a 1.27 mm pitch |
LCR meter | NF Corp. | ZM2376 | For measuring the AC properties of the coil |
Manipulator | NARISHIGE Group. | SM-15L | For manipulating the coil |
Medetomidine | Kobayashi Kako, Fukui, Japan | – | For anathesis of animals |
Midazolam | Astellas Pharma, Tokyo, Japan | – | For anathesis of animals |
Miniature screw | KOFUSEIBYO Co., Ltd. | S0.6*1.5 | For EEG-senseing and reference electrode |
Mouse | Japan SLC, Inc. | C57BL/6J (C57BL/6JJmsSlc) | Experimental animal |
Permalloy-45 rod | The Nilaco Corp. | 780544 | The core of the coil |
Recording system | Plexon Inc. | OmniPlex | For EEG data acquisition |
Stainless wire | Wakisangyo Co., Ltd. | HW-136 | For grasp by manipulator |
Stereotaxic apparatus | NARISHIGE Group. | SR-5M-HT | To fix a mouse head |
Surface-mount connector | Useconn Electronics Ltd. | PH127-2x10MG | For connector to mount on the flexible 2D array |
Testing equipment (LCR meter) | NF Corp. | ZM2372 | Contact check and impedance measurements |
White PLA filament | Zhejiang Flashforge 3D Technology Co., Ltd | PLA-F13 | The material used for 3D-printing the donut-shaped disks |
Xylocaine Jelly 2% | Sandoz Pharma Co., Ltd. | – | lidocaine hydrochloride |