Visual Evoked Potential Recordings in Mice Using a Dry Non-invasive Multi-channel Scalp EEG Sensor

Summary

We designed a dry-type 16 channel EEG sensor which is non-invasive, deformable, and re-usable. This paper describes the whole process from manufacturing the proposed EEG electrode to signal processing of visual evoked potential (VEP) signals measured on a mouse scalp using a dry non-invasive multi-channel EEG sensor.

Abstract

For scalp EEG research environments with laboratory mice, we designed a dry-type 16 channel EEG sensor which is non-invasive, deformable, and re-usable because of the plunger-spring-barrel structural facet and mechanical strengths resulting from metal materials. The whole process for acquiring the VEP responses in vivo from a mouse consists of four steps: (1) sensor assembly, (2) animal preparation, (3) VEP measurement, and (4) signal processing. This paper presents representative measurements of VEP responses from multiple mice with a submicro-voltage signal resolution and sub-hundred millisecond temporal resolution. Although the proposed method is safer and more convenient compared to other previously reported animal EEG acquiring methods, there are remaining issues including how to enhance the signal-to-noise ratio and how to apply this technique with freely moving animals. The proposed method utilizes easily available resources and shows a repetitive VEP response with a satisfactory signal quality. Therefore, this method could be utilized for longitudinal experimental studies and reliable translational research exploiting non-invasive paradigms.

Introduction

As the number of patients with senile degenerative brain diseases such as dementia, Alzheimer's, Parkinsonian syndromes, and stroke have increased with an aging population and an increasing life expectancy, the long-term societal burden of these diseases has also increased^1,2,3. In addition, most neurodevelopmental diseases, such as Schizophrenia and autism, are accompanied by cognitive and behavioral disorders which affect a patient's whole life^2,3,4. For this reason, researchers have been struggling to improve the diagnosis, prevention, pathological understanding, long-term observation, and treatment of brain diseases. However, problems remain stemming from the brain's complexity and unrevealed disease pathologies. Translational research may be a promising tool for identifying solutions because it enables the transfer of basic research to clinical applications within a shorter time frame, at lower cost, and with a higher success rate in neuroscience fields^5,6,7. Another goal of translational research is to examine applicability in human subjects, which requires non-invasive experimental approaches in animals that allow comparisons to the same method for humans. These conditions have led to several significant needs for developing non-invasive animal preparation methods. One method is electroencephalography (EEG), which reveals cortical brain connectivity and activity two-dimensionally with high temporal resolution, and which benefits from a non-invasive protocol. The event-related potential recording (ERP) is one of the typical experimental paradigms that utilize EEG.

Numerous previous studies employed non-invasive EEG methods for targeting humans subjects, whereas invasive methods, such as implant screws and pole type electrodes, have been used in animal studies^8,9,10,11,12. The signal quality and characteristics of these methods are significantly reliant on the invasiveness of the sensor placement. For successful translational research, Garner emphasized using the same conditions for animal study as those used for human research¹³. For basic research using animals, however, non-invasive EEG methodologies are not prevalent. A novel approach using a non-invasive scalp EEG sensor system focusing on laboratory mice would be a reliable and efficient tool for translational research that can be applied to the non-invasive paradigms for human beings, as well.

Numerous mouse EEG studies led the way by commercializing PCB (printed circuit board) based multi-channel electrodes^14,15,16. Although they adopted an invasive method, they had a restricted number of channels (3-8), which made it harder to observe large-scale brain dynamics. Furthermore, applications can be restricted by their invasiveness and high cost. In another research study, the KIST (Korea Institute of Science and Technology) developed a 40 channel polyimide-based thin-film electrode and attached it to a mouse's skull^17,18,19,20. This work acquired the highest number of mouse EEG channels. It was, however, mechanically weak and not easy to reuse; therefore, it was inappropriate for long-term observations, leading to a weakened signal, possibly caused by an immune reaction. Meanwhile, Troncoso and Mégevand acquired a sensory evoked potential (SEP) on rodents' skulls with thirty-two stainless steel electrodes secured by a perforated Poly(methyl methacrylate) (PMMA, acrylic glass) grid^21,22,23. Despite their high signal quality, the electrodes were mechanically flexible and tender; therefore, they had difficulties being applied to multiple experiments. In addition, this method was still minimally invasive. Although these methods provide good signal quality, the surface area of a mouse's skull is limited, therefore the number of electrodes is restricted using a stainless pole-type electrode. A number of previous EEG studies for mice showed several limitations. In this study, we will show a new method for measuring EEG applicable in pre-clinical translational research using a non-invasive dry-type multi-channel sensor.

In order to overcome the limitations of previous animal EEG methodologies, which included the intrinsic complexity of animal preparation, invasiveness, high cost, wastefulness, and weak mechanical strength, we sought to develop a new electrode which exhibits flexibility, dry type status, multi-channel capabilities, non-invasiveness, and re-usability. In the following protocol, we will describe the process of measuring visual evoked potential (VEP) recordings on a mouse scalp using a dry, non-invasive, multi-channel EEG sensor. This method utilizes easily available resources, therefore lowering the barrier to entry into animal experimentation in the biomedical engineering field.

Protocol

Animal care and handling followed the institutional guideline of the Gwangju Institute of Science and Technology (GIST).

NOTE: The procedure for acquiring the VEP signal from a mouse in vivo consists of four steps: (1) sensor assembly, (2) animal preparation, (3) VEP measurement, and (4) signal processing.

1. Sensor Assembly

1. Prepare sixteen pins for one non-invasive electrode.
   NOTE: Each pin-type electrode consists of three parts: a probe head plunger, an internal spring, and a barrel as shown in Figure 1a. The length of each pin is 13 mm, and the adjustable spring pre-load length is 1 mm.
2. Cut two pieces of glass fiber substrates (thickness: 1.5 mm) with the size of 15 mm × 17 mm (width × height).
   NOTE: The non-conducting fiberglass substrate functions as an insulator, which separates the multiple signals acquired simultaneously from the mouse's scalp.
3. Make sixteen holes of diameter 1.2 mm each using a precision engraving machine, as shown in Figure 1c.
   1. Spread out the probe coordination evenly at an interval of 2 mm onto the flat substrate: +7/0, +2/-2, +2/0, +2/+2, 0/-4, 0/-2, 0/0 (bregma), 0/+2, 0/+4, -2/-3, -2/-1, -2/+1, -2/+3, -4/-2, -4/0, -4/+2 (anteroposterior/lateral with basis of bregma, in mm)^24,25,26.
4. Stack two substrates and apply one drop of fast-acting adhesive glue between substrate layers, producing a double-layer of 3 mm thickness supporting sixteen stable and parallel electrodes during signal acquisition.
5. Assemble the sixteen electrodes onto the substrate one-by-one manually.
   NOTE: The smaller hole diameter stops each electrode at the same length. Each hole diameter is slightly smaller than the thickest diameter of a barrel within the single pin (1.3 mm) which enables the tight fixing of electrodes without any loosening.
6. Solder and link each electrode's ending solder-cup part to the touch proof connector.
7. Cover and hide the naked junctions with heat-shrink tubing for electrical insulation.

2. Animal Preparation

1. Anesthetize the mouse with an intraperitoneal (i.p.) injection of ketamine:xylazine 100:10 (100 mg/mL:10 mg/mL) mixture with the amount of 10 µL/g of body weight.
   NOTE: Check the animal's anesthesia is adequate by pulling one leg or tweaking the tail before starting preparation.
2. Apply eye ointment to keep the mouse's cornea moist with a cotton swab.
3. Remove the hairs around the head and shoulders with a hair-clipper, and then spread commercially available depilatory cream and keep it on this area for 3-4 min.
4. Remove the applied depilatory with a spatula, and then wipe up the rest with wet wipes applying water several times.

3. VEP Measurement

NOTE: The whole VEP measuring process took place in a dark Faraday cage (width × depth × height: 61 × 61 × 60 cm).

1. Mount the mouse's head onto the stereotaxic frame by placing ear bars into the mouse's ear canals and tightening them precisely in place.
2. Mount the sensor into the custom-made electrode holder (Figure 1b) and fix the sensor holder onto the stereotaxic frame, as shown in Figure 1d.
3. Locate the flexible EEG sensor, considering both the reference electrode position and the bregma position²⁷. After that, very carefully lower the sensor in the vertical direction so that the arrayed electrode plungers contact the mouse's scalp evenly onto the curved margin.
   NOTE: The lowered distance is smaller than 1 mm, which is the adjustable length of the plunger.
4. Check that the impedances are within the proper range from 100 kΩ to 2 MΩ. Reposition the electrode when any impedance value of the pin is out of the range²⁸.
5. Position the photo stimulator 20 cm away from the mouse's eyes.
6. Before starting the experiment, adapt the mouse for 10 min in the dark cage for dark visual adaptation.
   1. Set the parameters of the experimental devices as follows: Sampling frequency: 500 Hz; Notch filtering: 60 Hz; Inter-stimulus interval: 10 s; Flash duration: 10 ms; number of flash stimuli: 100 trials/subject.
      NOTE: The flash light is a white light LED illumination which has 550 ± 20 % lx with a distance of 20 cm.

4. VEP Responses Signal Processing Procedures

1. Epoching
   1. For continuously measuring serial data, extract each epoch to create single-trial VEP segments from the pre-stimulus period (-300 ms) to the post-stimulus period (600 ms), based on the flash stimulus onset.
      NOTE: Since we repeatedly provide flash stimulation over 100 trials for each subject, a total of 100 VEP epochs for each mouse are extracted in this step. EEG epoching is a process in which specific time-windows are extracted from the continuously measured EEG signal data.
2. Re-reference (average reference)
   1. Compute the average of EEG signals across all fourteen electrode channels at each time point and then subtract the averaged value from each channel. Repeat this procedure for all the VEP epochs.
3. Perform band-pass filtering of signal from ~1-100 Hz using a finite impulse response (FIR) filter.
4. Baseline correction
   1. Calculate the average of the EEG signals in the pre-stimulus period (baseline period, -300 ~ 0 ms) for each channel, then subtract this average from each point in the waveform (-300 ~ 600 ms). This adjusts the amplitude axis of VEP responses to facilitate the observation of brain wave changes after stimulation. Repeat this step for all of the VEP epochs.
5. Grand VEP Responses
   1. Average the single-trial VEP epochs to create single-subject averaged VEP waveforms for each channel. Then, calculate the grand ensemble average of VEP responses for each channel with respect to all the subjects.

Results

We calculated the ensemble average of VEP responses from eleven mice as shown in Figure 2. This result shows the VEP responses obtained through this experiment from the pre-stimulation period (-300 ms) to the post-stimulation period (600 ms), as the stimulation is given at time 0 s. It is noticeable that the signal fluctuates only for a while (less than 300 ms) after the stimulation, while the signal steadily stabilizes over time during the post-stimulation p...

Discussion

We firstly focused on the design of the sensor, prioritizing practicality by minimizing complex surgical procedures. The deformable EEG sensor is comprised of sixteen pins: fourteen for recording, one for ground, and the last one for reference electrodes. Each electrode has the plunger-spring-barrel structure, which applies deformability onto the electrode's contact surface, so they facilitate uniform and stable signal acquisition from the curved and tender mouse's scalp. Considering the welfare of the animals, we tried ...

Materials

Name	Company	Catalog Number	Comments
Ketamine 50 Inj. (Vial)	Yuhan	-	Ketamine HCl 57.68 mg
Zoletil 50 Inj.	Virbac	-	Tiletamina 125 mg/ Zolazepam 125 mg
Rompun 2% Inj.	BAYER	-	Xylazine hydrochloride 23.32mg/mL
Hycell solution 2%	Samil	-	Hydroxypropylmethylcellulose 20 mg
Puralube Vet Ointment 3.5 mg	Pharmaderm	-
Saline solution Inj.	JW Pharmaceutical	-	NaCl 9 g/1000 mL
Veet Hair Removal Cream – Legs & Body - Sensitive Skin	Reckitt Benckiser	-	depilatory
Skins - Surgical Skin Marker	Surgmed	S-3000	STERILE - Multi-Tip Fine Marker with ruler and label set
Stainless Steel Micro Spatulas	HEATHROW SCIENTIFIC	HS15907	One Round Flat End, 2L x 5/16W"
cotton swap
Stereotaxic, Desktop Digi Single	RWD Life Science	68025
Mouse Adapter	RWD Life Science	68010
Ear Bar for Mouse Non-Rupture	RWD Life Science	68306
Mitsar-EEG 202-24	MITSAR		amplifier
EEGStudio EEG acquisition software	MITSAR
White flash stimulator	MITSAR	MITSAR Flash stimulator
BCI2000 software	Schalk lab
g.USBamp	g.tec	0216
g.Power-g.USBamp	g.tec	0247
441 style straight body Touch Proof connector	PlasticsOne	441000PSW080001	441 - 000 PSW 80" (BLACK)
Standard probe	LEENO	SK100CSW	http://www.globalinterpark.com/detail/detail?prdNo=2114277241&dispNo=001851006012
Precision engraving machine tools	TINYROBO	TinyCNC-6060C
Heat shirink	3M	FP301